Portable Falling Weight Deflectometer Study15. Supplementary Notes N/A 16. Abstract: This research...
Transcript of Portable Falling Weight Deflectometer Study15. Supplementary Notes N/A 16. Abstract: This research...
PORTABLE FALLING WEIGHT DEFLECTOMETER STUDY
Byran C. Steinert, Dana N. Humphrey,
and Maureen A. Kestler
March 11, 2005
NETCR52 NETC Project No. 00-4
Prepared for New England Transportation Consortium
Prepared by: Department of Civil and Environmental Engineering
University of Maine Orono, Maine
with assistance from: USDA Forest Service
This report, prepared in cooperation with the New England Transportation Consortium, does not constitute a standard, specification or regulation. The contents of this report reflect the views of the authors who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the views of the New England Transportation Consortium or the Federal Highway Administration.
Technical Report Documentation Page1. Report No. 2. Government Accession No. 3. Recipient’s Catalog No.
NETCR52 N/A N/A 4. Title and Subtitle 5. Report Date
March 11, 2005 6. Performing Organization Code
Portable Falling Weight Deflectometer Study
N/A
7. Author(s) 8. Performing Organization Report No.
Bryan C. Steinert, Dana N. Humphrey, & Maureen A. Kestler
N/A
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
N/A 11. Contract or Grant No.
Department of Civil and Environmental Engineering University of Maine 5711 Boardman Hall Orono, ME 04469-5711
N/A 12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
FINAL 14. Sponsoring Agency Code
New England Transportation Consortium 179 Middle Turnpike Unit 5202 University of Connecticut Storrs, CT 06269-5202
N/A 15. Supplementary Notes
N/A 16. Abstract: This research investigated the effectiveness of the Portable Falling Weight Deflectometer (PFWD) for evaluating the support capacity of pavements during the spring thaw and evaluating the adequacy of subgrade and base compaction during construction. The performance of ten asphalt and gravel surfaced low volume roads were evaluated through spring thaw and recovery. Comparisons were made to the traditional FWD as well as other portable measuring devices. It was shown that the PFWD was able to follow seasonal stiffness variations and compared well with FWD derived moduli on both asphalt and gravel surfaces. Recommendations were made for using a PFWD to determine when to place and remove load restrictions. Field and laboratory tests were conducted to develop correlations between composite modulus, percent compaction, and water content for a range of aggregate types typical of New England. Comparisons were made between multiple PFWDs. A tentative technique was recommended for using a PFWD for compaction quality control for aggregate base and subbase courses. This is based on a rough equivalency between the PFWD composite modulus and percent compaction for aggregate at optimum water content. Factors are provided to correct the modulus at the field water content to the equivalent value at optimum.
17. Key Words 18. Distribution Statement
falling weight deflectometer, thaw weakening, load restrictions, low-volume roads, spring thaw, compaction, quality control, modulus, aggregate base course
No restrictions. This document is available to the public through the National Technical Information Service Springfield, Virginia 22161
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified 299 N/A
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
iii
PORTABLE FALLING WEIGHT DEFLECTOMETER STUDY NETC Project No. 00-4
EXECUTIVE SUMMARY
By: Bryan C. Steinert, Dana N. Humphrey, and Maureen A. Kestler
Department of Civil & Environmental Engineering, University of Maine, Orono, Maine
Portable falling weight deflectometers (PFWD) were investigated as a tool to aid in determining when to impose weight restrictions on low-volume roads during the spring thaw, and for compaction quality control for aggregate base courses and other soils. PFWDs operate on the same principle as conventional falling weight deflectometers (FWD), wherein a falling weight applies a force to a plate and the resulting deflection is measured using one or more deflection sensors. An advantage of a PFWD over a FWD is the former’s significantly lower purchase and operation costs. Comparing a PFWD to a nuclear moisture density meter (NDM) for compaction quality control, the latter has burdensome licensing and operational requirements due to the nuclear source. Moreover, PFWDs directly measure stiffness of pavement systems and compacted layers which is needed for mechanistic pavement design.
There are several pervious studies that compared composite modulus values for paved road determined by PFWDs, FWDs and Benkelman Beams. In general, the comparisons showed marginal correlation coefficients (r2) less than 0.5, however one study obtained an r2 of 0.86 for a correlation between moduli determined by the FWD and PFWD. The PFWD generally produced higher modulus values than the FWD, possibly due to the smaller depth of influence of the PFWD resulting in the stiff pavement having a greater influence on the resulting modulus. Some investigators imply that the PFWD is better suited to roads with thin pavements. Limited studies have been conducted to evaluate the PFWD as a tool for tracking seasonal stiffness variations. For this application, correlation coefficients relating the PFWD and the Benkelman Beam were generally high. For thin pavement sections, the PFWD adequately followed seasonal stiffness variations. Work done in Washington Sate has suggested that using deflection data to aid in load restriction placement and removal can be done and recommends that during the spring thaw, restrictions should be placed once the stiffness drops below 40 to 50% of their fully recovered values and then removed when the stiffness recovers to above these values.
A Prima 100 PFWD was selected as the primary instrument for this research because it can be used with three different drop weights, three plate diameters, adjustable fall heights, and up to three deflection sensors. Other PFWDs did not have this level of flexibility. Results from the Prima 100 PFWD were compared to other similar devices.
Spring Thaw Monitoring The performance of seven paved and three gravel surfaced roads were monitored
during the spring of 2004. Test sites were located in Maine, New Hampshire, and Vermont. One of the gravel surfaced sites located in New Hampshire was also monitored during the spring of 2003. Two additional sites in Northern Maine were used for testing
iv
on one day as part of an ongoing MaineDOT research project. Thermocouples, thermistors, and frost tubes were used at selected sites to monitor the advance and retreat of freezing conditions during late winter and spring months. Vibrating wire and standpipe piezometers were installed at selected sites to monitor pore water pressures in the subbase and subgrade layers. Time Domain Reflectometry (TDR) probes were used to monitor water content through the spring thaw and recovery periods at some sites. Instrumentation was used to examine the extent to which the road had thawed and provided the context for interpretation of PFWD and FWD results.
Prima 100 PFWD and traditional FWD measurements were taken at a minimum of eight locations at each test site. Measurements were taken approximately weekly during the spring thaw period. In addition, Loadman PFWD measurements were taken at spring thaw test sites in Rumney, New Hampshire. Clegg Impact Hammer and Humboldt Soil Stiffness Gauge measurements were taken at the United States Forest Service (USFS) Parking Lot during the spring of 2003 and 2004. With the Prima 100 PFWD, six measurements were taken at each of three different drop heights, at each test location. The first reading was neglected and the average of the remaining five was used for analysis and comparison. In addition, five Loadman PFWD, four Clegg Impact Hammer, and one Soil Stiffness Gauge measurement was taken at each test location. Moduli were backcalculated from FWD data using either DARWin or Evercalc.
Subsurface temperatures measurements taken at asphalt surfaced test sites indicated freezing temperatures penetrated to their maximum depths between February 17 and March 24, 2004. Maximum depths ranged from 866 mm (34 in.) to 1930 mm (76 in.). Complete thaw occurred at all test sites between mid-March and mid-April. Measurements taken at gravel surfaced test sites indicated freezing temperatures penetrated to their maximum depths between March 1 and April 21, 2004. Maximum depths ranged from 1128 mm (44 in.) to 2134 mm (84 in.). Complete thaw had occurred at all sites between early April and mid May. At most sites higher porewater pressures in the subgrade and subbase soils were associated with the thawing period. This is a factor that could contribute to reduction of pavement stiffness during the spring thaw.
For each test site, Prima 100 PFWD composite modulus, and where it is available, FWD asphalt, subbase, subgrade, and composite modulus and Loadman PFWD composite modulus values were plotted versus date. In general, for asphalt surfaced test sites, the moduli were high when the pavement section is frozen and during the early part of the thaw period when section is partially thawed. At some field sites there were significant differences in moduli from nearby test locations and from one week to the next. This variability is more apparent in gravel surfaced test sites compared to asphalt surfaced test sites. For both asphalt and gravel surfaced test sites, composite moduli generally decreased as thawing progressed. It was anticipated that a distinct minimum would occur before increasing through the recovery period. However, this was only evident at the Buffalo Road (NH), USFS Parking Lot (NH), Knapp Airport Parking Lot (VT), Crosstown Road (VT) , and to a lesser extent Stinson Lake Road (NH). At the remaining sites, the composite modulus that was reached during the spring thaw was about the same as, or in some, cases greater than the values measured during the summer. FWD derived layer moduli confirm these observations. In general, portable and traditional FWD moduli follow similar trends for both asphalt and gravel surfaced test
v
sites through the monitoring period. Thus, the PFWD and FWD would be equally as effective in monitoring stiffness change during the spring thaw.
The degree of correlation between composite moduli backcalculated using FWD and Prima 100 PFWD results were investigated. This was done for five sites in Maine where the composite moduli from the FWD were available. Regression analyses yielded correlation coefficients ranging from 0.336 to 0.950. In general, correlation coefficients tended to increase as pavement thickness decreased. The data from three test sites with asphalt thicknesses less than or equal to 127 mm (5 in.) were combined and produced the best correlation with r2 = 0.873. Two test sites with an asphalt thickness of 152 mm (6 in.) followed with r2 = 0.559. However, when excluding unreasonably high moduli greater than 4000 MPa, the correlation improves to r2 = 0.802. Route 1A (ME) was the single test site with a 180 mm (7 in.) asphalt thickness and produced the poorest correlation with r2 = 0.336. A regression analysis combining all asphalt surfaced test sites produced a correlation coefficient of 0.531. Again, when moduli greater than 4000 MPa are excluded, the correlation improved with r2 = 0.809. These results suggest that the PFWD could be used as an alternative to conventional FWDs for estimation of composite moduli of pavement sections for some pavement thicknesses.
Loadman and Prima 100 PFWD composite moduli were compared to FWD derived subbase moduli for two asphalt surfaced test sites in Rumney, New Hampshire. The Loadman PFWD provides a composite modulus that is less than that provided by the Prima 100. The Prima 100 PFWD correlates better to FWD derived subbase moduli (r2 = 0.552) than composite moduli obtained from the Loadman PFWD (r2 = 0.245).
The effect of Prima 100 PFWD drop weight (10, 15, and 20 kg (22, 33, and 44 lb)), plate diameters (100, 200, and 300 mm (4, 8, and 12 in.)), and drop height (850, 630, and 420 mm (33.5, 24.8, and 16.5 in.)) were investigated. In general, composite modulus increased with decreasing drop weight but was independent of plate diameter for the larger two drop weights. At the 10 kg (22 lb) drop weight the moduli decreased with increasing loading plate diameter. A possible explanation for this behavior is that a small plate diameter and drop weight influence only the upper portions of the pavement section and thus the deflection responses are dominated by the stiffer pavement layer, producing a larger composite modulus. In general, reduced drop heights produce moduli that are slightly less than moduli derived from using the full (850 mm) drop height. For most applications the largest drop weight, largest plate diameter, and largest drop height should be used for testing.
Up to three deflection sensors can be used with the Prima 100 PFWD. However, when backcalculating moduli current software makes use of only one sensor’s results at a time. Moduli derived from measurements from the outer two geophones are significantly greater than the composite moduli determined from the center geophone. Until software is developed to incorporate the deflections from all three geophones simultaneously into a backcalculation routine, the additional geophones provide little useful additional information.
Six Prima 100 PFWD measurements were taken at each of three different drop heights, at each test location. For the majority of points tested at the field sites; the first measurement was less than subsequent measurements. This was consistent with
vi
observations made by other researchers. On the average, first drop was less than the average of the remaining five drops by nearly 10%. However, the second drop was only 1% less than the average of the remaining four drops. This shows that the results of the first drop should always be neglected. It is recommended that the results from drops two through six be averaged to obtain results that are representative of a test location.
Field testing techniques for monitoring seasonal stiffness variation in paved and unpaved low volume roads using the Prima 100 PFWD were developed. The core of the recommendations is that load restrictions are placed once the composite moduli measured with the PFWD drops below 80% of the fully recovered baseline value measured during the summer and early fall. The load restriction is then removed when the moduli recover to 80% of the baseline value. The selection of 80% is arbitrary since the amount of damage that would occur at the reduced modulus depends on individual pavement sections, allowable vehicle weight, and traffic levels. Assessment of these factors was beyond the scope of this study. Baseline and spring thaw measurements should be made at the same locations. During the early portion of thawing period, it may be necessary to take daily readings to monitor the sometimes rapid decrease in composite modulus.
Compaction Control
Five field test sites in Maine, New Hampshire, and Connecticut were used to evaluate the effectiveness of the PFWD as a tool to monitor compaction. Tests were performed at a minimum of 12 locations, utilizing both the Prima 100 PFWD and Nuclear Moisture Density Meter (NDM). Samples were taken at each site for sieve analysis, maximum dry density, and optimum water content determination. The field component included tests on two subgrades, one construction sand, two base aggregates, and one reclaimed stabilized base product.
The primary purpose of the laboratory component of this project was to determine a relationship between PFWD results and percent compaction under controlled conditions. Tests were performed on five soil types representative of New England base and subbase aggregates. These materials included: one crushed material, one construction sand, and three base/subbase aggregates. The tests were conducted in a 1.8 m x 1.8 m x 0.9 m (6 ft x 6 ft x 3 ft) deep test container. Material was added to the container in approximately 152 mm (6 in.) lifts. Each lift was compacted using a hand tamper and electric jackhammer with a modified flat plate attachment. Each aggregate was compacted in the container to approximately 90, 95, and 100% of the maximum dry density (AASHTO T 180). The effect of water content was determined at 95% of the maximum dry density. Measurements were taken at optimum water content as well as ± 3% of the optimum water content. Once all the material was compacted in the test container, the following portable testing devices were used: Prima 100 PFWD, Clegg Impact Hammer, NDM, and Dynamic Cone Penetrometer (DCP). Prima 100 PFWD and Clegg Impact Hammer measurements were taken in the same manner as was done for the spring thaw portion of the research. In addition, one sand cone test was completed, and two water content samples were taken for each trial for comparison to NDM measurements. NDM measurements were used as the prime basis for comparison.
For the laboratory tests, the composite moduli generally increased as percent compaction increased. This was true for all samples with the exception of the New
vii
Hampshire Gravel which exhibited the opposite trend. With the exception of the New Hampshire Sand, the correlation coefficients were less than 0.5 indicating poor correlation. Combining all the results yielded a correlation coefficient of 0.045, indicating no correlation. However, including only the results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel resulted in a higher correlation coefficient of 0.35, but still indicating a poor correlation. Results from the field test sites also indicate that as the degree of compaction increases, composite modulus increases. In general, correlation coefficients were greater for field test results compared to laboratory test results. Combining the results for the three base materials tested in the field, resulted in a correlation coefficient of 0.818, which is a relatively strong correlation. However, the significance of this correlation is diminished by the fact the water content at all the field sites was dry of optimum.
Laboratory results also show that there is a general trend that the composite moduli tends to decrease as water content increases. Correlation coefficients ranged from 0.003 (Connecticut Crushed Gravel) to 0.814 (Wardwell Gravel). The low correlation coefficients for several of the samples are due in part to the role that percent compaction plays in the composite modulus, which is not accounted for when only water content is considered. Combining at the laboratory results yielded a correlation coefficient of 0.285 which indicates poor correlation. For measurements taken at field sites the correlation coefficient ranged from 0.008 (Route 25 Gravel) to 0.521 (Route 25 Sand). However, water contents measured at field sites were generally drier than -3% of the OWC and in some instances were as low as -9%, which are significantly different from water contents obtained during laboratory tests.
Multivariable linear regression analyses were used to determine the best fit line for composite modulus as a function of percent compaction and water content. The coefficient of multiple determination (R2) for the laboratory materials ranged from 0.141 (Connecticut crushed gravel) to 0.867 (Wardwell gravel). Combining all laboratory samples produced an R2 of 0.326. However, including only laboratory results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel increased the R2 to 0.624. This indicates that 62% of the variation in composite modulus is explained by the percent compaction and water content relative to optimum. The R2 for the field materials ranged from 0.001 (Route 25 gravel) to 0.679 (I-84 crushed gravel). Combining the three field sites where granular base was tested yielded an R2 of 0.823, which indicates a reasonably strong correlation of composite modulus with percent compaction and water content, independent of the type of material tested. However, the water contents for the field sites were all dry of optimum which may limit the significance of this result. The multi-variable linear regressions based on the three laboratory samples indicated above and field results yielded predicted composite modulus at 95% percent compaction that agreed within 20% which is reasonable agreement. Overall, the analysis shows that both percent compaction and water content relative to optimum have an important influence on composite modulus.
Based on the results of this research the tentative procedure given below is recommended for using the Prima 100 PFWD to monitor compaction of granular base courses. The procedure is based on the observation that there is a rough equivalency between percent compaction and composite modulus for granular base at optimum water
viii
content. Correction factors are recommended to correct the composite modulus measured at the field water content to the equivalent value at optimum water content. The regression equation for the combined laboratory results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel was used to derive the recommendations. This equation was used since it had a higher correlation coefficient than the regression that included all five laboratory samples combined and it had a larger range of water contents than the field samples.
The target composite modulus at optimum water content should be chosen based on Table 1 that gives a rough equivalency with percent compaction based on AASTHO T-180. Composite moduli measured in the field should be corrected to the equivalent composite modulus at optimum water content by adding the factors given in Table 2. Thus, it is necessary to determine the field water content relative to OWC to apply this procedure. Possibilities for measuring the water content include oven drying, pan drying, Speedy Moisture Meter©, time domain reflectometry, or nuclear density meter in backscatter mode. The researchers caution that the values given in Tables 1 and 2 are based on a limited dataset. It is recommended that these equivalences be confirmed for additional materials used by individual state DOTs.
Table 1 Tentative equivalences between percent compaction and composite modulus at optimum water content for base and subbase course aggregate.
Percent Compaction based onAASTHO T-180 (%)
Equivalent Prima 100 PFWDComposite Modulus (MPa) at
Optimum Water Content 90 92 95 115 98 130 100 139
Table 2 Factor to correct composite modulus measured at field water content to equivalent value at optimum water content.
Water Content Relative to Optimum
Correction Factor to be Added to Composite
Modulus (MPa) Measured at Field Water Content
-4% -31 -3% -23 -2% -15 D
ry o
f O
WC
-1% -8 At OWC 0
+1% 8 +2% 15 +3% 23 W
et o
f O
WC
+4% 31
ix
PORTABLE FALLING WEIGHT DEFLECTOMETER STUDY
Byran C. Steinert, Dana N. Humphrey, and Maureen A. Kestler
Prepared for New England Transportation Consortium
March 11, 2005
NETCR52 NETC Project No. 00-4
Prepared by: Department of Civil and Environmental Engineering
University of Maine Orono, Maine
with assistance from: USDA Forest Service
This report, prepared in cooperation with the New England Transportation Consortium, does not constitute a standard, specification or regulation. The contents of this report reflect the views of the authors who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the views of the New England Transportation Consortium or the Federal Highway Administration.
x
ACKNOWLEDGEMENTS
The researchers acknowledge the New England Transportation Consortium for
providing funding for this research project. We especially thank Bill Real, Steve Colson,
Chris Benda, Don Larsen, Matt Turo, and James Walls who served on the New England
Transportation Consortium Technical Committee assigned to this project. Additional
thanks go to Charlie Smith from the Cold Regions Research & Engineering Laboratory
and to Jim Raymond of the Vermont Agency of Transportation.
xi
TABLE OF CONTENTS
EXECUTIVE SUMMARY ............................................................................................. iii
ACKNOWLEDGEMENTS ............................................................................................. x
TABLE OF CONTENTS ................................................................................................ xi
LIST OF TABLES ......................................................................................................... xiv
LIST OF FIGURES ..................................................................................................... xviii
1. INTRODUCTION......................................................................................................... 1 1.1 BACKGROUND ...................................................................................................... 1
1.2 SCOPE OF STUDY.................................................................................................. 3
1.3 ORGANIZATION OF THE REPORT..................................................................... 3
2. LITERATURE REVIEW ............................................................................................ 5
2.1 INTRODUCTION .................................................................................................... 5
2.2 PFWD AS ALTERNATIVE FOR COMPACTION CONTROL ............................ 5 2.2.1 Current Compaction Control Methods................................................... 5 2.2.2 PFWD Equipment ................................................................................ 10 2.2.3 Past Test Programs, Results, and Recommendations .......................... 18
2.3 PFWD AS TOOL TO EVALUATE THAW WEAKENING OF ROADS............ 40 2.3.1 Current Methods to Evaluate Thaw Weakening of Roads................... 42 2.3.2 Past Test Programs, Results, and Recommendations .......................... 50
2.4 PFWD QUESTIONNAIRE .................................................................................... 51 2.4.1 2003 Results......................................................................................... 52 2.4.2 2004 Results......................................................................................... 52
2.5 SUMMARY............................................................................................................ 52
3. FIELD & LABORATORY TEST PROTOCOL ..................................................... 55 3.1 INTRODUCTION .................................................................................................. 55
3.2 FIELD TEST SITE LOCATIONS ......................................................................... 55 3.2.1 Seasonally Posted Low Volume Roads ............................................... 55 3.2.2 Compaction Control Field Test Sites ................................................... 76
3.3 INSTRUMENTATION .......................................................................................... 83 3.3.1 Frost Penetration Measurement ........................................................... 83 3.3.2 Pore Water Pressure Measurement ...................................................... 90
3.4 FIELD TESTING PROCEDURES ........................................................................ 95 3.4.1 Spring Thaw Monitoring...................................................................... 95
xii
3.4.2 Subgrades and Construction Materials .............................................. 104
3.5 LABORATORY TESTING PROCEDURES ...................................................... 107 3.5.1 Spring Thaw Monitoring.................................................................... 107 3.5.2 Subgrades and Construction Materials .............................................. 107
3.6 SUMMARY.......................................................................................................... 111
4. SPRING THAW MONITORING ........................................................................... 113
4.1 INTRODUCTION ................................................................................................ 113
4.2 FROST PENETRATION ..................................................................................... 114
4.3 PORE WATER PRESSURE ................................................................................ 116
4.4 SEASONAL STIFFNESS VARIATIONS........................................................... 122 4.4.1 Backcalculation of Layer Moduli ...................................................... 122 4.4.2 Asphalt Surfaced Roads..................................................................... 127 4.4.3 Gravel Surfaced Roads ...................................................................... 136
4.5 COMPARISON OF PFWD AND FWD MODULI ............................................. 142 4.5.1 Route 11 & Route 167 Field Test Sites ............................................. 142 4.5.2 Composite Modulus ........................................................................... 147 4.5.3 Subbase Modulus ............................................................................... 154
4.6 COMPARISON TO OTHER PORTABLE DEVICES........................................ 169
4.7 EVALUATION OF FIELD TESTING TECHNIQUES ...................................... 171 4.7.1 Loading Plate Diameter and Drop Weight......................................... 171 4.7.2 Drop Height ....................................................................................... 173 4.7.3 Moduli Derived from Additional Geophones .................................... 176 4.7.4 Multiple Measurements at Each Test Location ................................. 179
4.8 RECOMMENDATIONS...................................................................................... 181 4.8.1 Factors that Affect Need for Seasonal Load Restrictions .................. 182 4.8.2 Field Testing Techniques ................................................................... 182 4.8.3 Application of Procedure to Field Sites ............................................. 185
4.9 SUMMARY.......................................................................................................... 189
5. COMPACTION CONTROL ................................................................................... 193 5.1 INTRODUCTION ................................................................................................ 193
5.2 IN-PLACE WATER CONTENT AND DRY DENSITY .................................... 194
5.3 FACTORS AFFECTING COMPOSITE MODULUS......................................... 199 5.3.1 Effect of Percent Compaction on Composite Modulus ..................... 200 5.3.2 Effect of Water Content on Composite Modulus............................... 217 5.3.3 Multivariate Linear Regression.......................................................... 226 5.3.4 Additional Factors Influencing Composite Modulus......................... 231
5.4 COMPARISON OF PORTABLE DEVICES....................................................... 232 5.4.1 Prima 100 PFWD Comparison .......................................................... 234
xiii
5.4.2 Clegg Impact Hammer ....................................................................... 238 5.4.3 Effect of Operator Technique ............................................................ 241
5.5 RECOMMENDATIONS...................................................................................... 242
5.6 SUMMARY.......................................................................................................... 246
6. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ......................... 251 6.1 SUMMARY.......................................................................................................... 251
6.1.1 Literature Review............................................................................... 253 6.1.2 Field & Laboratory Test Protocol ...................................................... 255 6.1.3 Spring Thaw Monitoring.................................................................... 258 6.1.4 Compaction Control........................................................................... 265
6.2 CONCLUSIONS .................................................................................................. 269 6.2.1 Spring Thaw Monitoring.................................................................... 270 6.2.2 Field Testing Techniques ................................................................... 270 6.2.3 Compaction Control........................................................................... 271
6.3 RECOMMENDATIONS FOR FURTHER RESEARCH.................................... 272
REFERENCES.............................................................................................................. 273
APPENDIX A - GRAIN SIZE DISTRIBUTION CURVES .................................... 279
APPENDIX B - MOISTURE DENSITY CURVES................................................. 289
APPENDIX C - COMPACTION CONTROL RAW LABORATORY DATA ..... 295
xiv
LIST OF TABLES
Table 2.1 Correlation coefficients of 20 test points on two TMS structures (SHT, 1998). .................................................................................................23
Table 2.2 Summary of correlations between the FWD and GDP, TFT, and Prima 100 PFWD at Mountsorrel and Bardon test sites (Fleming, et al., 2000). ......................................................................................................33
Table 2.3 Summary of correlations between the Prima 100 PFWD and GDP and TFT at Mountsorrel and Bardon test sites (Fleming, et al., 2000).........34
Table 2.4 Description of foundations applied to theoretical model (Thom and Fleming, 2002)..............................................................................................35
Table 2.5 Predicted surface moduli from different dynamic plate test devices (Thom and Fleming, 2002). ..........................................................................35
Table 2.6 Road user feedback to DOTs and USFS on spring thaw load (Kestler, et al., 2000). ...................................................................................44
Table 2.7 Benefits from seasonal load restrictions (FHWA, 1990)..............................45
Table 2.8 Summary of recommendations made by Kestler after Rutherford, et al. (1985).......................................................................................................48
Table 3.1 Summary of seasonally posted low volume road field test sites. .................57
Table 3.2 Laboratory properties of in-situ material at Kennebec Road, Hampden/Dixmont, Maine. ..........................................................................59
Table 3.3 Laboratory properties of in-situ material at Lakeside Landing Road, Glenburn, Maine. ..........................................................................................60
Table 3.4 Laboratory properties of in situ subbase material at Stinson Lake Road, Rumney, New Hampshire. .................................................................62
Table 3.5 Laboratory properties of in situ material at Buffalo Road, Rumney, New Hampshire. ...........................................................................................63
Table 3.6 Laboratory properties of in situ material at Crosstown Road, Berlin, Vermont. .......................................................................................................65
Table 3.7 Laboratory properties of in situ material at Knapp Airport Parking Lot, Berlin, Vermont.....................................................................................66
Table 3.8 Laboratory index properties of cohesive subgrade material at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000).....................................69
xv
Table 3.9 Laboratory index properties of subgrade material at Route 126, Monmouth/Litchfield, Maine (Helstrom and Humphrey, 2005). .................72
Table 3.10 Water contents at thermocouple location on Route 1A Frankfort/Winterport, Maine (Fetten and Humphrey, 1998)........................75
Table 3.11 Test Section description of Route 11, Wallagrass Plantation, Maine (Bouchedid and Humphrey, 2004)................................................................76
Table 3.12 Test Section description of Route 167, Presque Isle/Fort Fairfield, Maine (Bouchedid and Humphrey, 2004). ...................................................76
Table 3.13 Summary of compaction control field test sites............................................77
Table 3.14 Summary of Marshall Stability tests on field samples at Commercial Paving & Recycling test site, Scarborough, Maine. .....................................82
Table 3.15 Summary of instruments spring thaw field test sites ....................................83
Table 3.16 Summary of thermocouple locations at spring thaw field test sites..............85
Table 3.17 Summary of thermistor locations at spring thaw field test sites ...................86
Table 3.18 Summary of frost tube locations at spring thaw field test sites ....................87
Table 3.19 Summary of vibrating wire piezometer locations at spring thaw field test sites.........................................................................................................91
Table 3.20 Summary of standpipe piezometer locations at spring thaw field test sites ...............................................................................................................93
Table 3.21 TDR probe locations at Stinson Lake Road and USFS Parking Lot, Rumney, New Hampshire.............................................................................94
Table 3.22 Summary of test point locations at Witter Farm Road, Route 126, and Route 1A ................................................................................................97
Table 3.23 Prima 100 PFWD input parameters ..............................................................99
Table 3.24 FLEX testing plan drop sequence used at Berlin, Vermont test sites (LTTP, 2000). .............................................................................................103
Table 3.25 FLEX testing plan target loads used at Berlin, Vermont test sites (LTTP, 2000). .............................................................................................104
Table 4.1 Summary of frost penetration measurements made on asphalt surfaced test sites. .......................................................................................115
xvi
Table 4.2 Summary of frost penetration measurements made on gravel surfaced test sites. .......................................................................................116
Table 4.3 Summary of standpipe piezometer measurements......................................119
Table 4.4 Summary of time domain reflectometry probe water content readings.......................................................................................................119
Table 4.5 Summary of manual vibrating wire piezometer measurements..................120
Table 4.6 Summary of PFWD and FWD composite moduli at the end of thawing and during recovery periods..........................................................128
Table 4.7 FWD and PFWD mean composite moduli for different asphalt thicknesses. .................................................................................................148
Table 4.8 Summary of the effects of reduced drop height on PFWD composite modulus for different asphalt thicknesses...................................................174
Table 4.9 Summary of the effects of reduced drop height on PFWD composite moduli for different asphalt thicknesses. ....................................................177
Table 4.10 Comparison of the first, second, and third measurements with successive measurements at Route 126, Monmouth/Litchfield, Maine. .........................................................................................................181
Table 4.11 Prima 100 PFWD input parameters. ...........................................................184
Table 4.12 Summary of load restrictions for spring thaw field test sites......................186
Table 5.1 Summary of laboratory samples .................................................................195
Table 5.2 Summary of laboratory measurements. ......................................................196
Table 5.3 Summary of field samples ..........................................................................199
Table 5.4 Summary of the correlations between percent compaction and composite modulus for laboratory samples. ...............................................204
Table 5.5 Summary of the correlations between percent compaction and composite modulus for field samples. ........................................................212
Table 5.6 Summary of the correlations between water content and composite modulus for laboratory samples..................................................................217
Table 5.7 Summary of the correlations between water content and composite modulus for field samples...........................................................................222
xvii
Table 5.8 Summary of multivariate linear regression analyses on laboratory samples........................................................................................................227
Table 5.9 Summary of multivariate linear regression analyses on field tests on granular base. ..............................................................................................227
Table 5.10 Comparison of average composite moduli associated with varying degrees of compaction. ...............................................................................231
Table 5.11 Comparison of composite moduli from USFS and UMaine Prima 100 PFWDs.................................................................................................238
Table 5.12 Comparison of Prima 100 PFWD composite moduli and CIH moduli ......239
Table 5.13 Comparison of Prima 100 PFWD composite moduli determined by different users. ............................................................................................242
Table 5.14 Tentative equivalences between percent compaction and composite modulus at optimum water content for base and subbase course aggregate.....................................................................................................244
Table 5.15 Factor to correct composite modulus measured at field water content to equivalent value at optimum water content. ...........................................244
Table 5.16 Prima 100 PFWD input parameters. ...........................................................245
Table C.1 Summary of Connecticut crushed gravel raw laboratory data....................295
Table C.2 Summary of New Hampshire sand raw laboratory data. ............................296
Table C.3 Summary of New Hampshire gravel raw laboratory data. .........................297
Table C.4 Summary of Owen J. Folsom gravel raw laboratory data. .........................298
Table C.5 Summary of Wardwell gravel raw laboratory data.....................................299
xviii
LIST OF FIGURES
Figure 2.1 Nuclear density and water content determination: (a) direct transmission; (b) backscatter; (c) air gap (Holtz and Kovacs, 1981)..............8
Figure 2.2 Keros Prima 100 PFWD. ..............................................................................12
Figure 2.3 Keros Prima 100 PFWD. ..............................................................................12
Figure 2.4 Profile of load cell and geophone. ................................................................13
Figure 2.5 Bottom view of geophone.............................................................................13
Figure 2.6 Additional geophones. ..................................................................................14
Figure 2.7 Prima 100 PFWD PDA display after one measurement. ..............................15
Figure 2.8 Loadman PFWD (Livneh, et al., 1997).........................................................16
Figure 2.9 Loadman PFWD display after one measurement. ........................................17
Figure 2.10 Comparison of Loadman and FWD at various test points (Gros, 1993). ............................................................................................................19
Figure 2.11 Correlation between Loadman and FWD (Gros, 1993)................................19
Figure 2.12 Comparison of Loadman and FWD at various test points (Gros, 1993). ............................................................................................................20
Figure 2.13 Correlation between Loadman and FWD (Gros, 1993)................................20
Figure 2.14 Comparison of Loadman and FWD at various test points (Whaley, 1994). ............................................................................................................21
Figure 2.15 Correlation between Loadman and FWD (Whaley, 1994). ..........................21
Figure 2.16 Comparisons of Loadman, FWD, and Benkelman Beam (SHT, 1998). ............................................................................................................23
Figure 2.17 (a) Uncorrected and (b) corrected relationship between FWD modulus and PFWD modulus using the double testing technique (Livneh, 1997). .............................................................................................25
Figure 2.18 Loadman PFWD deflection versus Benkelman Beam deflection (Livneh, et al., 1998).....................................................................................26
Figure 2.19 Comparison of multiple FWD and PFWD measurements at one location (Honkanen, 1991). ..........................................................................26
xix
Figure 2.20 Comparison of Loadman and FWD at various test points (Gros, 1993). ............................................................................................................28
Figure 2.21 Correlation between Loadman and FWD (Whaley, 1994). ..........................28
Figure 2.22 Comparison of FWD, Loadman PFWD, Benkelman Beam, and Clegg Hammer at various test points (Whaley, 1994)..................................29
Figure 2.23 Moduli versus location for granular base material (Siekmeier, et al., 2000). ............................................................................................................30
Figure 2.24 Correlation between Loadman and FWD moduli (Pidwerbesky, 1997). ............................................................................................................32
Figure 2.25 Relationship between stiffness modulus determined by the portable dynamic plate test devices and the FWD (on subgrade and 400 mm thick granular capping) (Fleming, et al., 2000). ...........................................34
Figure 2.26 Comparison of k value from Prima 100 and HFWD at multiple test locations (Kamiura, et al., 2000)...................................................................36
Figure 2.27 Comparison of the variation in deflection ratio with the number of drops at one location (Kamiura, et al., 2000)................................................37
Figure 2.28 Prima 100 PFWD measured increase in stiffness due to increased compactive effort and time (Nazzal, 2003). .................................................39
Figure 2.29 Correlations between Prima 100 PFWD, FWD, and PLT (Nazzal, 2003). ............................................................................................................39
Figure 2.30 Typical signage associated with placing load restrictions (Janoo and Cortez, 1998). ...............................................................................................41
Figure 2.31 Methods for determining when to place and remove load restrictions (Kestler, et al., 2000). ...................................................................................43
Figure 2.32 Length of time over which load restrictions are placed (Kestler, et al., 2000). ......................................................................................................44
Figure 2.33 Comparison in the ability of the Loadman PFWD and Benkelman Beam to track strength change through spring thaw (Davies, 1997)............51
Figure 3.1 Typical condition of Kennebec Road, Hampden/Dixmont, Maine in April, 2003. ...................................................................................................56
Figure 3.2 Approximate geographic location of spring thaw test sites. .........................58
Figure 3.3 Typical road condition of Lakeside Landing Road, Glenburn, Maine in April, 2003................................................................................................60
xx
Figure 3.4 Typical road condition of Stinson Lake Road, Rumney, New Hampshire in July, 2003. ..............................................................................61
Figure 3.5 Existing site conditions at USFS Parking Lot, Rumney, New Hampshire.....................................................................................................64
Figure 3.6 Existing conditions at Crosstown Road, Berlin, Vermont in March 2004 ..............................................................................................................65
Figure 3.7 Plan view of test sections at Witter Farm Road (Lawrence, et al., 2000). ............................................................................................................67
Figure 3.8 Cross section of test sections at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000). ...............................................................................68
Figure 3.9 Test section layout of Route 126 Monmouth/Litchfield, Maine (Helstrom and Humphrey, 2005). .................................................................71
Figure 3.10 Test section layout of Route 1A Frankfort/Winterport, Maine (Fetten and Humphrey, 1998). ..................................................................................74
Figure 3.11 I-84 test section, Southington, Connecticut. .................................................77
Figure 3.12 Route 25 Test Section 1, Effingham/Freedom, New Hampshire..................79
Figure 3.13 Route 25 Test Section 2, Effingham/Freedom, New Hampshire..................79
Figure 3.14 Route 26 test section, New Gloucester, Maine. ............................................80
Figure 3.15 Thermocouple string detail ...........................................................................84
Figure 3.16 Frost tube detail.............................................................................................87
Figure 3.17 Thermocouple/thermistor section view.........................................................88
Figure 3.18 Placement of thermocouple string in auger hole...........................................89
Figure 3.19 Standpipe piezometer detail. .........................................................................92
Figure 3.20 Standpipe piezometer detail for Witter Farm Road, Orono, Maine (Lawrence, et al., 2000). ...............................................................................92
Figure 3.21 Kennebec Road and Lakeside Landing Road test point layout. ...................96
Figure 3.22 Stinson Lake Road, Buffalo Road, and Crosstown Road test point layout ............................................................................................................96
Figure 3.23 Variable drop heights (a) 850 mm, (b) 630 mm, and (c) 420 mm................98
xxi
Figure 3.24 MaineDOT JILS Model 20 C FWD............................................................101
Figure 3.25 CRREL Dynatest 8000 FWD......................................................................102
Figure 3.26 VAOT Dynatest 8000 FWD. ......................................................................103
Figure 3.27 Test point layout for compaction control field test sites.............................105
Figure 3.28 CPN MC-1 Portaprobe NDM .....................................................................106
Figure 3.29 Troxler 3430 NDM. ....................................................................................106
Figure 3.30 Laboratory test box. ....................................................................................108
Figure 3.31 Bosch 11304 hammer with modified flat plate attachment ........................109
Figure 3.32 Laboratory test point layout ........................................................................110
Figure 4.1 Formation of ice lenses within a pavement structure (WSDOT)................118
Figure 4.2 Route 126 (Section 3), Monmouth/Litchfield, Maine automated pore water pressure measurements. ....................................................................121
Figure 4.3 Route 126 (Section 8), Monmouth/Litchfield, Maine automated pore water pressure measurements. ....................................................................121
Figure 4.4 Evercalc 5.0 general file data entry screen (WSDOT, 2001). ....................125
Figure 4.5 Evercalc 5.0 raw FWD data conversion screen (WSDOT, 2001). .............126
Figure 4.6 Evercalc 5.0 FWD deflection data file screen (WSDOT, 2001).................126
Figure 4.7 Stiffness variation at Kennebec Road (Section 1), Hampden/Dixmont, Maine. ........................................................................129
Figure 4.8 Stiffness variation at Kennebec Road (Section 2), Hampden/Dixmont, Maine. ........................................................................130
Figure 4.9 Stiffness variation at Buffalo Road, Rumney, New Hampshire .................130
Figure 4.10 Stiffness variation at Stinson Lake Road, Rumney, New Hampshire ........131
Figure 4.11 Stiffness variation at Knapp Airport Parking Lot, Berlin, Vermont...........131
Figure 4.12 Stiffness variation at Witter Farm Road (Control Section), Orono, Maine ..........................................................................................................132
Figure 4.13 Stiffness variation at Witter Farm Road (Section 2), Orono, Maine. .........132
xxii
Figure 4.14 Stiffness variation at Witter Farm Road (Section 1), Orono, Maine. .........133
Figure 4.15 Stiffness variation at Route 126 (Section 3), Monmouth/Litchfield, Maine ..........................................................................................................133
Figure 4.16 Stiffness variation at Route 126 (Section 8), Monmouth/Litchfield, Maine ..........................................................................................................134
Figure 4.17 Stiffness variation at Route 126 (Section 12), Monmouth/Litchfield, Maine ..........................................................................................................134
Figure 4.18 Stiffness variation at Route 1A (Section D-1), Frankfort/Winterport, Maine ..........................................................................................................135
Figure 4.19 Stiffness variation at Route 1A (Section D-2), Frankfort/Winterport, Maine ..........................................................................................................135
Figure 4.20 Stiffness variation at Route 1A (Section D-3), Frankfort/Winterport, Maine ..........................................................................................................136
Figure 4.21 Stiffness variation at Lakeside Landing Road (Section 1), Glenburn, Maine ..........................................................................................................137
Figure 4.22 Detailed stiffness variation at Lakeside Landing Road (Section 1), Glenburn, Maine .........................................................................................138
Figure 4.23 Stiffness variation at Lakeside Landing Road (Section 2), Glenburn, Maine ..........................................................................................................138
Figure 4.24 Detailed stiffness variation at Lakeside Landing Road (Section 2), Glenburn, Maine .........................................................................................139
Figure 4.25 2003 stiffness variation at USFS Parking Lot, Rumney, New Hampshire...................................................................................................139
Figure 4.26 2004 stiffness variation at USFS Parking Lot, Rumney, New Hampshire...................................................................................................140
Figure 4.27 2004 detailed stiffness variation at USFS Parking Lot, Rumney, New Hampshire...................................................................................................140
Figure 4.28 Stiffness variation at Crosstown Road, Berlin, Vermont............................141
Figure 4.29 Detailed stiffness variation at Crosstown Road, Berlin, Vermont..............141
Figure 4.30 Modulus versus test location at Route 11 (Test Pit 1), Wallagrass Plantation, Maine ........................................................................................143
xxiii
Figure 4.31 Modulus versus test location at Route 11 (Test Pit 2), Wallagrass Plantation, Maine ........................................................................................143
Figure 4.32 Modulus versus test location at Route 11 (Test Pit 3), Wallagrass Plantation, Maine ........................................................................................144
Figure 4.33 Modulus versus test location at Route 11 (Test Pit 4), Wallagrass Plantation, Maine ........................................................................................144
Figure 4.34 Modulus versus test location at Route 167 (Test Pit 1), Presque Isle/Fort Fairfield, Maine............................................................................145
Figure 4.35 Modulus versus test location at Route 167 (Test Pit 2), Presque Isle/Fort Fairfield, Maine............................................................................145
Figure 4.36 Modulus versus test location at Route 167 (Test Pit 3), Presque Isle/Fort Fairfield, Maine............................................................................146
Figure 4.37 Modulus versus test location at Route 167 (Test Pit 4), Presque Isle/Fort Fairfield, Maine............................................................................146
Figure 4.38 Comparison of FWD and PFWD composite moduli at Kennebec Road, Hampden/Dixmont, Maine...............................................................148
Figure 4.39 Comparison of FWD and PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine......................................................................149
Figure 4.40 Comparison of FWD and PFWD composite moduli at Witter Farm Road, Orono, Maine....................................................................................149
Figure 4.41 Comparison of FWD and PFWD composite moduli at Route 11, Wallagrass Plantation, Maine .....................................................................150
Figure 4.42 Comparison of FWD and PFWD composite moduli at Route 167, Presque Isle/Fort Fairfield, Maine ..............................................................150
Figure 4.43 Comparison of FWD and PFWD composite moduli at Route 1A, Frankfort/Winterport, Maine ......................................................................151
Figure 4.44 Comparison of FWD and PFWD composite moduli for asphalt thicknesses ≤ 127 mm (5 in.). .....................................................................151
Figure 4.45 Comparison of FWD and PFWD composite moduli for asphalt thicknesses equal to 152 mm (6 in.). ..........................................................152
Figure 4.46 Comparison of FWD and PFWD composite moduli for asphalt thicknesses equal to 152 mm (6 in.) and moduli ≤ 4000 MPa....................152
xxiv
Figure 4.47 Comparison of FWD and PFWD composite moduli for all asphalt surfaced test sites ........................................................................................153
Figure 4.48 Comparison of FWD and PFWD composite moduli for all asphalt surfaced test sites and moduli ≤ 4000 MPa ................................................153
Figure 4.49 Comparison of FWD and PFWD composite moduli at Lakeside Landing Road, Glenburn, Maine. ...............................................................154
Figure 4.50 Comparison of FWD subbase moduli and PFWD composite moduli at Kennebec Road, Hampden/Dixmont, Maine ..........................................155
Figure 4.51 Comparison of FWD subbase moduli and PFWD composite moduli at Stinson Lake Road, Rumney, New Hampshire ......................................156
Figure 4.52 Comparison of FWD subbase moduli and PFWD composite moduli at Buffalo Road, Rumney, New Hampshire ...............................................156
Figure 4.53 Comparison of FWD subbase moduli and PFWD composite moduli at Knapp Airport Parking Lot, Berlin, Vermont.........................................157
Figure 4.54 Comparison of FWD subbase moduli and PFWD composite moduli at Witter Farm Road, Orono, Maine ...........................................................157
Figure 4.55 Comparison of FWD subbase moduli and PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine ............................................158
Figure 4.56 Comparison of FWD subbase moduli versus PFWD composite moduli at Route 1A, Frankfort/Winterport, Maine.....................................158
Figure 4.57 Comparison of FWD subbase moduli and PFWD composite moduli for asphalt thicknesses ≤ 127 mm (5 in.). ................................................159
Figure 4.58 Comparison of FWD subbase moduli and PFWD composite moduli for asphalt thickness ≤ 127 mm (5 in.) and moduli ≤ 5000 MPa ...............159
Figure 4.59 Comparison of FWD subbase moduli and PFWD composite moduli for all asphalt surfaced test sites .................................................................160
Figure 4.60 Comparison of FWD subbase moduli and PFWD composite moduli for all asphalt surfaced test sites and moduli ≤ 5000 MPa .........................160
Figure 4.61 Comparison of FWD subbase moduli and PFWD composite moduli at Crosstown Road, Berlin, Vermont..........................................................161
Figure 4.62 Comparison of FWD and PFWD ISM at Kennebec Road, Hampden/Dixmont, Maine .........................................................................162
xxv
Figure 4.63 Comparison of FWD and PFWD ISM at Stinson Lake Road, Rumney, New Hampshire...........................................................................163
Figure 4.64 Comparison of FWD and PFWD ISM at Buffalo Road, Rumney, New Hampshire ..........................................................................................163
Figure 4.65 Comparison of FWD and PFWD ISM at Knapp Airport Parking Lot, Berlin, Vermont ..........................................................................................164
Figure 4.66 Comparison of FWD and PFWD ISM at Witter Farm Road, Orono, Maine ..........................................................................................................164
Figure 4.67 Comparison of FWD and PFWD ISM at Route 126, Monmouth/Litchfield, Maine......................................................................165
Figure 4.68 Comparison of FWD and PFWD ISM at Route 1A, Frankfort/Winterport, Maine ......................................................................165
Figure 4.69 Comparison of FWD and PFWD ISM for test sites with asphalt thicknesses ≤ 127 mm (5 in.). .....................................................................166
Figure 4.70 Comparison of FWD and PFWD ISM for test sites with asphalt thicknesses equal to 152 mm (6 in.). ..........................................................166
Figure 4.71 Comparison of FWD and PFWD ISM for all asphalt surfaced test sites .............................................................................................................167
Figure 4.72 Comparison of FWD and PFWD ISM at Lakeside Landing Road, Glenburn, Maine .........................................................................................167
Figure 4.73 Comparison of FWD and PFWD ISM at Crosstown Road, Berlin, Vermont ......................................................................................................168
Figure 4.74 Comparison of FWD and PFWD ISM at USFS Parking Lot, Rumney, New Hampshire...........................................................................168
Figure 4.75 Comparison of FWD and PFWD ISM for all gravel surfaced test sites .............................................................................................................169
Figure 4.76 Comparison of FWD derived subbase moduli to Loadman and Prima 100 PFWD composite moduli on asphalt surfaced test sites ......................171
Figure 4.77 Effect of drop weight and loading plate diameter on Prima 100 PFWD composite moduli............................................................................172
Figure 4.78 Effect of drop height on PFWD composite moduli at Witter Farm Road, Orono, Maine....................................................................................174
xxvi
Figure 4.79 Effect of drop height on PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine......................................................................175
Figure 4.80 Effect of drop height on PFWD composite moduli at Route 1A, Frankfort/Winterport, Maine ......................................................................175
Figure 4.81 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Kennebec Road, Hampden/Dixmont, Maine .........................................................................177
Figure 4.82 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Route 126, Monmouth/Litchfield, Maine......................................................................178
Figure 4.83 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Route 1A, Frankfort/Winterport, Maine ..........................................................................................................178
Figure 4.84 Effect of consecutive drops on composite modulus on May 12, 2004 at Route 126 (Section 3), Monmouth/Litchfield, Maine ...................179
Figure 4.85 Effect of consecutive drops on composite modulus values on April 22, 2004 at Route 126 (Section 12), Monmouth/Litchfield, Maine ...........180
Figure 4.86 Buffalo Road, Rumney, New Hampshire. ..................................................187
Figure 4.87 Knapp Airport Parking Lot, Berlin, Vermont. ............................................187
Figure 4.88 USFS Parking Lot, Rumney, New Hampshire. ..........................................188
Figure 4.89 Crosstown Road, Berlin, Vermont..............................................................188
Figure 5.1 Comparison of oven dried and NDM water contents. ................................197
Figure 5.2 Comparison of percent compaction determined from sand cone and NDM tests. ..................................................................................................198
Figure 5.3 Effect of percent compaction on composite modulus, Connecticut crushed gravel. ............................................................................................200
Figure 5.4 Effect of percent compaction on composite modulus, New Hampshire sand...........................................................................................201
Figure 5.5 Effect of percent compaction on composite modulus, New Hampshire gravel........................................................................................201
Figure 5.6 Effect of percent compaction on composite modulus, OJF gravel. ............202
xxvii
Figure 5.7 Effect of percent compaction on composite modulus, Wardwell gravel. .........................................................................................................202
Figure 5.8 Comparison of percent compaction and composite modulus, Connecticut crushed gravel.........................................................................204
Figure 5.9 Comparison of percent compaction and composite modulus, New Hampshire sand...........................................................................................205
Figure 5.10 Comparison of percent compaction and composite modulus, New Hampshire gravel........................................................................................205
Figure 5.11 Comparison of percent compaction and composite modulus, OJF gravel. .........................................................................................................206
Figure 5.12 Comparison of percent compaction and composite modulus, Wardwell gravel..........................................................................................206
Figure 5.13 Comparison of percent compaction and composite modulus for all laboratory samples. .....................................................................................207
Figure 5.14 Comparison of percent compaction and composite modulus for three laboratory samples. .....................................................................................207
Figure 5.15 Comparison of percent compaction and composite modulus for laboratory tests with water contents dry of the OWC.................................208
Figure 5.16 Comparison of percent compaction and composite modulus for selected laboratory samples with water contents dry of the OWC. ............209
Figure 5.17 Comparison of percent compaction and composite modulus for laboratory tests with water contents wet of the OWC. ...............................209
Figure 5.18 Comparison of percent compaction and composite modulus for selected laboratory samples with water contents wet of the OWC.............210
Figure 5.19 Comparison of percent compaction and composite modulus of crushed gravel tested at I-84, Southington, Connecticut. ...........................213
Figure 5.20 Comparison of dry density and composite modulus of subgrade tested at I-84, Southington, Connecticut.....................................................213
Figure 5.21 Comparison of percent compaction and composite modulus of construction sand tested at Route 25, Effingham/Freedom, New Hampshire...................................................................................................214
Figure 5.22 Comparison of percent compaction and composite modulus of gravel tested at Route 25, Effingham/Freedom, New Hampshire. ........................214
xxviii
Figure 5.23 Comparison of percent compaction and composite modulus of subgrade tested at CPR, Scarborough, Maine.............................................215
Figure 5.24 Comparison of percent compaction and composite modulus for materials tested at Route 25 and I-84 field test sites...................................215
Figure 5.25 Change in moduli with time at CPR test site, Scarborough, Maine. ..........216
Figure 5.26 Change in moduli with time at Route 201 test site, The Forks, Maine.......216
Figure 5.27 Comparison of water content and composite modulus, Connecticut crushed gravel. ............................................................................................218
Figure 5.28 Comparison of water content and composite modulus, New Hampshire sand...........................................................................................218
Figure 5.29 Comparison of water content and composite modulus, New Hampshire gravel........................................................................................219
Figure 5.30 Comparison of water content and composite modulus, OJF gravel. ..........219
Figure 5.31 Comparison of water content and composite modulus, Wardwell gravel. .........................................................................................................220
Figure 5.32 Comparison of water content and composite modulus for all laboratory samples. .....................................................................................220
Figure 5.33 Prima 100 PFWD measurement on Wardwell gravel wet of optimum ......221
Figure 5.34 Prima 100 PFWD measurement on Wardwell gravel wet of optimum ......221
Figure 5.35 Comparison of water content and composite modulus of crushed gravel tested at I-84, Southington, Connecticut..........................................223
Figure 5.36 Comparison of water content and composite modulus of subgrade tested at I-84, Southington, Connecticut.....................................................223
Figure 5.37 Comparison of water content and composite modulus of sand tested at Route 25, Effingham/Freedom, New Hampshire. ..................................224
Figure 5.38 Comparison of water content and composite modulus of gravel tested at Route 25, Effingham/Freedom, New Hampshire. ........................224
Figure 5.39 Comparison of water content and composite modulus of subgrade tested at CPR, Scarborough, Maine. ...........................................................225
Figure 5.40 Comparison of water content and composite modulus for all field test sites.......................................................................................................225
xxix
Figure 5.41 Composite modulus predicted by regression equations at 4% dry of optimum. .....................................................................................................229
Figure 5.42 Composite modulus predicted by regression equations at optimum. .........229
Figure 5.43 Composite modulus predicted by regression equations at 4% wet of optimum. .....................................................................................................230
Figure 5.44 Effect of percent compaction and water content on predicted composite modulus based on laboratory results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel. ....................230
Figure 5.45 Typical shapes of coarse grained bulky particles (Holtz and Kovacs, 1981). ..........................................................................................................233
Figure 5.46 Poorly graded New Hampshire sand...........................................................233
Figure 5.47 Well graded Connecticut crushed gravel. ...................................................234
Figure 5.48 Testing with USFS and UMaine Prima 100 PFWDs..................................235
Figure 5.49 Change in composite modulus with subsequent drops for two PFWDs at the same test point for OJF gravel at 100% compaction and optimum water content (TP #1). ..........................................................236
Figure 5.50 Change in composite modulus with subsequent drops for two PFWDs at the same test point for Wardwell gravel at 100% compaction and optimum water content (TP #1)........................................237
Figure 5.51 Comparison of USFS and UMaine Prima 100 PFWD composite moduli. ........................................................................................................237
Figure 5.52 Change in composite modulus with subsequent drops for different devices at the same test point for OJF gravel at 100% compaction and optimum water content (TP #1). ..........................................................239
Figure 5.53 Clegg Impact Hammer measurement on New Hampshire sand. ................240
Figure 5.54 Comparison of Clegg Impact Hammer and UMaine Prima 100 PFWD composite moduli............................................................................240
Figure 5.55 MaineDOT representative (operator #2) performing laboratory PFWD measurements. ................................................................................242
Figure A.1 Grain size distribution of Connecticut crushed gravel................................279
xxx
Figure A.2 Grain size distribution of New Hampshire sand. ........................................279
Figure A.3 Grain size distribution of New Hampshire gravel. .....................................280
Figure A.4 Grain size distribution of bottom 1 ft of OJF gravel...................................280
Figure A.5 Grain size distribution of Owen J. Folsom gravel. .....................................281
Figure A.6 Grain size distribution of Wardwell gravel.................................................281
Figure A.7 Grain size distribution of crushed gravel tested at I-84, Southington, Connecticut. ................................................................................................282
Figure A.8 Grain size distribution of existing subgrade material tested at I-84, Southington, Connecticut. ..........................................................................282
Figure A.9 Grain size distribution of construction sand tested at Route 25, Effingham/Freedom, New Hampshire........................................................283
Figure A.10 Grain size distribution of gravel tested at Route 25, Effingham/Freedom, New Hampshire........................................................283
Figure A.11 Grain size distribution of MaineDOT Type D gravel tested at Route 26, New Gloucester, Maine. .......................................................................284
Figure A.12 Grain size distribution of MaineDOT Type E gravel tested at Route 26, New Gloucester, Maine. .......................................................................284
Figure A.13 Grain size distribution of existing subgrade tested at CPR, Scarborough, Maine....................................................................................285
Figure A.14 Grain size distribution of existing subgrade material at Route 126 (Section 3), Monmouth/Litchfield, Maine..................................................285
Figure A.15 Grain size distribution of existing subgrade material at Route 126 (Section 8), Monmouth/Litchfield, Maine..................................................286
Figure A.16 Grain size distribution of existing subbase material at Stinson Lake Road, Rumney, New Hampshire. ...............................................................286
Figure A.17 Grain size distribution of tire chip / soil mixtures at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000). ...........................................287
Figure A.18 Grain size distribution of MaineDOT Type D subbase used at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000). ..................................287
Figure A.19 Grain size distribution of MaineDOT Type D subbase used at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000). ..................................288
xxxi
Figure A.20 Grain size distribution of subbase material at Route 1A, Frankfort/Winterport, Maine (Fetten and Humphrey, 1998)......................288
Figure B.1 Connecticut crushed gravel moisture density curve. ..................................289
Figure B.2 New Hampshire sand moisture density curve.............................................289
Figure B.3 New Hampshire gravel moisture density curve. .........................................290
Figure B.4 Owen J. Folsom #1 moisture density curve. ...............................................290
Figure B.5 Owen J. Folsom #2 moisture density curve. ...............................................291
Figure B.6 Wardwell gravel moisture density curve. ...................................................291
Figure B.7 Moisture density curve of sand tested at Route 25, Effingham/Freedom, New Hampshire........................................................292
Figure B.8 Moisture density curve of gravel tested at Route 25, Effingham/Freedom, New Hampshire........................................................292
Figure B.9 Moisture density curve of MaineDOT Type D at Route 26, New Gloucester, Maine.......................................................................................293
Figure B.10 Moisture density curve of MaineDOT Type E at Route 26, New Gloucester, Maine.......................................................................................293
Figure B.11 Moisture density curve of existing subgrade material at CPR, Scarborough, Maine....................................................................................294
xxxii
(BLANK PAGE)
1
CHAPTER 1
INTRODUCTION
1.1 BACKGROUND
Modulus (stiffness) is one of the primary inputs into mechanistic pavement design
procedures and provides insight into long term pavement performance. Based on layer
stiffness and thickness, stresses and strains are computed to investigate whether they are
below critical limits needed to achieve adequate pavement performance during the design
life (Van Gurp, et al., 2000). Despite the importance of modulus, some aspects of
pavement construction and management are still based on measurement of parameters
that are not directly connected with long-term performance or on empirical based
judgments. Two critical areas that do not make use of modulus are evaluating the support
capacity of pavements during the spring thaw for the purposes of restricting truck loading
and evaluating the adequacy of subgrade and base compaction during construction.
These topics are the focus of this study.
Pavements in areas with seasonal freezing and thawing often undergo frost heave
and thaw weakening in addition to load-induced pavement distress. To minimize
damage, many road maintenance agencies impose load restrictions during damage-
susceptible periods. Spring thaw adversely affects pavement life while load restrictions
impose local economic hardships throughout the northern United States and Canada.
Although the maximum allowable load and the duration of the reduced load period vary
widely among agencies, they try to strike a balance between minimizing the disruption to
the local economy caused by the load restrictions and minimizing road damage.
2
Although modulus is a key parameter in determining damage-susceptibility of
pavements, the imposition of spring-thaw load restrictions are often based on visual
observation combined with the pavement manager’s judgment. Modulus could be
monitored during spring thaw and through recovery using a trailer-mounted falling
weight deflectometers (FWD). However, FWD purchase, operation, and maintenance is
expensive. Second, even if a state owns a FWD, it can only cover so many roads within
the spring thaw period. As a result, determining when the road has thawed and recovered
sufficient strength to remove the restriction is left to personal experience and subjective
judgment.
Virtually all state highway departments use dry density as the principal criterion
for judging the quality of compacted earthwork. This criterion implies that increased dry
density produces improved engineering properties in the material. Although the use of
dry density for field control can be easily accomplished, particularly with the increasing
use of nuclear devices, its value as a usable criterion is only valid insofar as the dry
density does, in fact, indicate the critical engineering properties of the material such as
stiffness (Langfelder and Nivargikar, 1967). At present, there is no viable alternative to
density as a method for compaction control since there is no well-established method to
measure the stiffness of compacted materials that would be practical to use during
construction.
The purpose of this project is to investigate a practical method of measuring
stiffness of pavement structures during the spring thaw and of compacted subgrade soil
and base aggregate during construction. The premise is that this can be accomplished
using a portable falling weight deflectometer (PFWD). The PFWD operates on the same
3
principle as the full-size FWD, but it is small enough that it can be easily moved by one
person.
1.2 SCOPE OF STUDY
There are two main objectives for this research project. The first objective was to
investigate the ability of portable falling weight deflectometers (PFWD) to track seasonal
stiffness variations. Measurements were taken on paved and unpaved low volume roads
during the spring thaw. Comparisons were made to the traditional FWD as well as other
portable devices. Correlations were developed to compare performance.
Recommendations were made for field testing techniques.
The second objective was to investigate the ability of the PFWDs to serve as an
alternative to traditional compaction control devices. A relationship between PFWD
composite moduli and percent compaction for soil types representative of New England
base and subbase aggregates was established. The effect of water content was also
investigated. Comparisons were made to other portable devices and correlations were
developed. Recommendations were made for field testing techniques.
1.3 ORGANIZATION OF THE REPORT
This report is divided into six chapters, and is organized as follows. Chapter 2 is
a literature review covering several topics relevant to the evaluation of the PFWD.
PFWD and other portable testing devices are described. Past test programs, results, and
recommendations for use on both asphalt and gravel surfaces are presented. Current
methods to evaluate thaw weakening of roads are discussed. Results and
4
recommendations for using the PFWD to track seasonal stiffness variations are made for
asphalt surfaced test sites. A questionnaire to determine current usage of the PFWD as an
alternative to traditional compaction control devices and as a tool to evaluate thaw
weakening of roads was distributed to transportation agencies, the results of which are
presented.
Chapter 3 describes the field and laboratory testing techniques. Field test site
locations are presented including: current road condition, subsurface conditions, and
cross sections of the test sites. Instrumentation descriptions, and installation and
monitoring procedures are presented. Descriptions of the field testing procedures and
data gathered for both the study of seasonally posted low volume paved and unpaved
roads and the field and laboratory study of the compaction of subgrades and construction
materials are provided. Finally, data analysis techniques that were employed are
discussed.
Chapter 4 presents the analysis, field test results, and recommendations for
utilizing the PFWD as tool to track seasonal stiffness variations in paved and unpaved
low volume roads.
Chapter 5 presents the analysis, laboratory and field test results, and
recommendations for utilizing the PFWD as an alternative to traditional compaction
control devices.
Chapter 6 summarizes all aspects of the research project and provides conclusions
and recommendations for future work.
5
CHAPTER 2
LITERATURE REVIEW
2.1 INTRODUCTION
This literature review focuses on the use of a Portable Falling Weight
Deflectometer (PFWD) as an alternative to traditional compaction control methods and as
a tool to evaluate thaw weakening of roads. Several studies have been conducted
pertaining to the use, reliability, and accuracy of the PFWD. The PFWD has been
compared to traditional in situ testing devices and is discussed below. A questionnaire
was developed and distributed to state departments of transportation as well as
international organizations currently utilizing PFWDs, the results of which are also
discussed.
2.2 PFWD AS ALTERNATIVE FOR COMPACTION CONTROL
2.2.1 Current Compaction Control Methods
Traditional compaction control methods can be costly, require considerable time
for field tests, and some have extensive training and safety requirements. Devices used to
monitor compaction measure one of two things, density or modulus (stiffness). Stiffness
is a qualitative term meaning a general resistance to deformation. It is often used
interchangeably with elastic modulus, modulus of subgrade reaction, and resilient
modulus. It largely determines the strains and displacements of the subgrade as it is
loaded and unloaded (Newcomb and Birgisson, 1999). During construction major
emphasis has traditionally been placed on achieving the specified dry density and little
6
consideration is given to the engineering properties desired of the compacted fill. Dry
density and water content correlate well with the engineering properties, and thus they are
convenient construction control parameters (Holtz and Kovacs, 1981). Therefore,
instruments that measure density are used for quality assurance. However, modulus is
one of the primary inputs into any pavement design procedure, and provides insight into
long term pavement performance. Based on layer stiffness and thickness, critical stresses
and strains are computed to investigate whether they are below critical limits needed to
achieve adequate pavement performance during the design life (Van Gurp, et al., 2000).
Such stiffness measurements are more fundamentally sound from an engineering
perspective than the now universally accepted moisture-density measurements used for
compaction control (Lenke, et al., 2003). Stiffness measurements on pavement structures
make it possible to treat it in the same manner as other civil engineering structures by
using mechanistically based design methods. Selecting the type of rehabilitation to be
implemented on a given pavement is of considerable economic significance. To reach
that decision with inadequate knowledge of the structural condition of the pavement may
result in unnecessary costly repairs (Dynatest, 2004).
2.2.1.1 Density Measurement
Many in-situ testing devices exist that monitor density. Until recently, the sand
cone method (AASHTO T 191) and to a lesser extent the rubber balloon method
(AASHTO T 205) were used to measure density and provide quality control for
construction. However, both methods are labor intensive, time is required to dig holes,
and there is time delay in obtaining water content. In practice, both of these devices have
7
been replaced by the nuclear moisture density meter (NDM) (AASHTO T 238). The
NDM emits gamma rays and neutrons from two radioactive materials (Cesium-137 and
Americium-241) housed within the device. The gamma rays emitted by the Cesium-137
penetrate the layers through a probe inserted into the ground from beneath the unit and
interact with electrons in the material. A Geiger Detector housed within the unit counts
the gamma rays that reach it from the source. A calibration chart, provided by the
manufacturer, relates the count to density. Similarly, Americium-241 produces high
energy (fast) neutrons that collide with nuclei in the material. The neutrons that collide
with hydrogen nuclei slow down much quicker than those that collide with other, larger
nuclei. The Geiger Detector counts only low energy (slow) neutrons, thus the detector
count is proportional to the number of hydrogen atoms present. Since water (H2O)
contains many hydrogen atoms, the detector count is proportional to moisture content. A
calibration chart, provided by the manufacturer relates the count to moisture content.
Operation of the device is shown in Figure 2.1.
Nuclear tests can be conducted rapidly as compared with destructive methods,
with results known within minutes. Such rapidity allows the contractor and field
engineer to know results quickly, allowing corrective action to be taken as necessary
before additional earthwork has been placed (Lenke, et al., 2003). However, the NDM
has several drawbacks:
1. Relatively high initial cost.
2. Operation is limited to certified operators.
8
Figure 2.1 Nuclear density and water content determination: (a) direct transmission; (b) backscatter; (c) air gap (Holtz and Kovacs, 1981).
9
3. Those working with and around the device are required to wear film badges that
must periodically be submitted and checked for radioactive exposure dosage in
response to operational manuals, women of child-bearing age are frequently
discouraged by their supervisors from using this equipment.
4. Shipping and transport of the device over borders and state lines requires
significant monetary investment, paperwork, and therefore time.
5. Disposal at the end of the gauge’s useful life is extremely difficult.
2.2.1.2 Alternative Portable Device Measurement
It is for the reasons given above that a new method for evaluating compaction is
desirable. The disadvantages have prompted transportation agencies to look for
nonnuclear methods for compaction control. Such alternatives must eliminate the safety
and regulatory concerns of nuclear methods yet provide comparable speed and precision
during field testing. Any alternative must also provide an engineering measurement that
is related to the engineering properties and engineering performance of the soil evaluated
(Lenke, et al., 2003).
There are several portable test devices that attempt to measure the in situ modulus
of highway foundation material (Fleming, et al., 2002). The Clegg Impact Tester (ASTM
D 5874) utilizes a hammer of known mass dropped from a predetermined height to
evaluate the modulus. An accelerometer is used to measure the deceleration of the
weight as it impacts the underlying material, and is reported on a digital readout as an
10
impact value that may be correlated to stiffness (Thompson and Garcia, 2003). The
Dynamic Cone Penetrometer (DCP) test measures the penetration resistance as the cone
of the device is driven into the pavement structure. A cone on the bottom of an anvil is
driven into the ground by means of a hammer dropped from a standard height. The
amount of penetration per blow is monitored. The penetration is a function of the shear
strength of the material. The Humboldt Soil Stiffness Gauge (SSG) is a recently
developed nondestructive testing device that measures the in situ stiffness. The SSG
imparts very small displacements to the soil at 25 steady state frequencies between 100
and 196 Hz (Thompson and Garcia, 2003). It then measures the stress imparted to the
surface and resulting surface velocity as a function of time, from which, stiffness may be
calculated. The final type of device is the Portable Falling Weight Deflectometer
(PFWD), and is discussed in detail in the following sections.
2.2.2 PFWD Equipment
The PFWD is a light, portable device that has been developed to measure stiffness
of construction layers including subgrades, base courses, and pavements. Various models
have been developed and used significantly in Europe. The PFWD creates a non-
destructive shock-wave through the soil as a result of the impact of a falling weight.
Sensors such as velocity transducers or accelerometers are used to measure surface
movement, from which deflection is determined. A load cell is used to measure the
impact force of the falling weight. Boussinesq developed equations for the state of stress
within a homogeneous, isotropic, linearly elastic half-space for a point load acting
11
perpendicular to the surface (Holtz and Kovacs, 1981). Manipulation of this theory
provides a means for determining the modulus from the two measurements.
2.2.2.1 Prima 100 PFWD
The first model is the Prima 100, manufactured by Keros Technology and Carl
Bro Pavement Consultants, both of Denmark. This model is versatile in that it allows
more flexibility in the stress applied to the underlying material by varying the falling
mass (10, 15, and 20 kg (22, 33, and 44 lb)) as well as varying the drop height (10 to 850
mm (0.4 to 33.5 in.)). Three different loading plates may be used, 100, 200, and 300 mm
(3.9, 7.9, and 11.8 in.). The Prima 100 has a load impulse of between 15 and 20 ms and a
load range of 1 to 15 kN (224.8 to 3372.1 lbf), i.e. up to 200 kPa (29 psi) with its 300 mm
(11.8 in.) diameter bearing plate (Fleming, et al., 2000). The unit is shown in Figures 2.2
and 2.3.
The Prima 100 uses two types of sensors: a load cell for measuring the impact
force from the falling weight, and a geophone that measures the velocity of the surface
from which deflection is determined by integration (Christensen, 2003). With this model,
the reaction of the soil to the shock-waves can be measured by up to three geophones that
extend radially outward from the unit. The load cell and geophones are shown in Figures
2.4, 2.5, and 2.6.
12
Figure 2.2 Keros Prima 100 PFWD.
Figure 2.3 Keros Prima 100 PFWD.
13
Figure 2.4 Profile of load cell and geophone.
Figure 2.5 Bottom view of geophone.
14
Figure 2.6 Additional geophones.
The Prima 100 PFWD records values of force, pressure, and deflection with
respect to time and is recorded automatically by a Personal Digital Assistant (PDA).
Based on these measurements, the software calculates the elastic modulus according to
Equation 2.1. For modulus values at distances of more than two radii from the center of
the load may be determined by Equation 2.2 (Christensen, 2003). A sample of the output
screen for one measurement is shown in Figure 2.7.
0
02 )1(
davf
E⋅⋅−⋅
=σ
Eqn. 2.1
Where: E = Surface Modulus μ = Poisson’s ratio (default: 0.5) σ0 = Applied stress at surface a = Radius of loading plate d0 = Deflection f = Factor that depends on the stress distribution
15
Uniform: f = 2 (default) Rigid plate: f = π/2 Parabolic, granular: f = 8/3 Parabolic, cohesive: f = 4/3
)(
)1(
0
20
2
rdrav
E⋅
⋅⋅−=
σ Eqn. 2.2
Where: r = Distance from center d0(r) = Deflection at the distance r from the center
Figure 2.7 Prima 100 PFWD PDA display after one measurement.
2.2.2.2 Loadman PFWD
The Loadman PFWD was originally developed in Finland by Al-Engineering Oy
(Honkanen, 1991) for the testing of granular base courses and its use has been extended
to bound layers. It has been adopted by over 60 research organizations, universities,
consultants, road agencies, contractors and local authorities in Canada, Estonia, Finland,
16
India, Israel, Italy, Pakistan, Russia, Sweden, the United Kingdom and other countries
located throughout the world (Pidwerbesky, 1997). It is not as versatile as the Prima 100
model. The device utilizes a single 10 kg weight that is dropped from a fixed height of
800 mm (2.6 ft). The Loadman has loading plate sizes of 132, 200, and 300mm (5.2, 7.9,
and 11.8 in.). The device is believed to be capable of measuring deflections ranging from
0 to 5 mm (0 to 0.2 in.), with an approximate time of loading of between 25 and 30
milliseconds (ms) and maximum dynamic load of roughly 23 kN (5171 lbf) (Gros, 1993).
The Loadman PFWD is shown in Figure 2.8.
Figure 2.8 Loadman PFWD (Livneh, et al., 1997).
The Loadman PFWD uses two types of sensors, a load cell and an accelerometer.
Manipulation of Boussinesq’s equations provides an equation to determine modulus from
the results measured by the Loadman (Pidwerbesky, 1997). Equation 2.3 is used by the
device to determine elastic modulus.
17
Δ
×=paE 5.1 Eqn. 2.3
Where: Δ = deflection under Loadman plate p = unit load on circular plate a = radius of base plate E = modulus of elasticity
For each measurement, the Loadman displays the maximum deflection, calculated
bearing capacity modulus, length of the loading impulse, percentage of the rebound
deflection compared to the maximum deflection, and the compaction ratio, which is
defined as the deflections measured on second and subsequent drops divided by the
deflection measured on the first drop (Al-Engineering Oy). A sample of the output
screen after one measurement is shown in Figure 2.9.
Figure 2.9 Loadman PFWD display after one measurement.
18
2.2.3 Past Test Programs, Results, and Recommendations
Several test programs have been conducted outside of the United States (US) over
the last several years with an increasing but still limited number of studies within the US.
Generally, the primary purpose of the test programs was to determine whether the PFWD
was suitable for use as a measuring instrument for quality assurance purposes in road
construction. Most researchers developed relationships between the PFWD and other,
more traditional measuring devices. The studies yielded encouraging results. Testing
was conducted on pavement and unbound aggregate layers and each are discussed
separately below.
2.2.3.1 Pavement Layers
PFWD evaluations on pavement layers have been conducted by several
investigators. Flexible pavements and thin membrane surfaced roads were tested by Gros
(1993), Livneh (1997), Whaley (1994), and Davies (1997). Correlations to other
traditional measures of stiffness and comparison between PFWD devices have been made
and are discussed.
Gros (1993) performed tests on pavement sections consisting of asphalt concrete
(AC), thickness unknown, underlain by 0 to 64 mm (0 to 2.5 in.) of crushed gravel or
crushed rock, and 0 to 32 mm (0 to 1.25 in.) of gravel. Multiple Loadman PFWD and
Falling Weight Deflectometer (FWD) measurements were taken. Gros (1993)
recommended that when testing asphalt pavement, two measurements at each measuring
point are necessary. For the majority of points tested, the Loadman produced higher
modulus values than the FWD. Correlation coefficients between devices range from 0.03
to 0.44 with an average of 0.27. Typical results are shown in Figures 2.10 through 2.13.
19
Figure 2.10 Comparison of Loadman and FWD at various test points (Gros, 1993).
Figure 2.11 Correlation between Loadman and FWD (Gros, 1993).
Results of testing on bound layers indicated that the PFWD and FWD correlate poorly
with one another, typically differing by 20% to 30%. Gros attributes the “irregularity” of
the Loadman curve to the fact that it is more sensitive to heterogeneous layers because its
depth of influence is less. He notes that larger stone particles can cause peaks, and the
influence of a stone will be higher for Loadman because the “spheres” of stress levels are
smaller.
20
Figure 2.12 Comparison of Loadman and FWD at various test points (Gros, 1993).
Figure 2.13 Correlation between Loadman and FWD (Gros, 1993).
Similarly, Whaley (1994) directly compared the results of the Loadman PFWD
with the FWD to determine its effectiveness when testing deflection and layer moduli of
AC layers. The tested pavement section consisted of an 80 mm (3.2 in.) asphalt concrete
surface, 200 mm (7.9 in.) base course, 1220 mm (48 in.) subbase, and a rigid layer at
1500 mm (59 in.) depth (Whaley, 1994). Like Gros (1993), Whaley (1994) concluded
that the Loadman yielded higher moduli than the modulus values backcalculated from the
FWD measurements at all test points, with a difference of roughly 200 MPa between
21
FWD and Loadman results. The correlation coefficient was 0.2. The overall results of
the testing are shown in Figures 2.14 and 2.15.
Figure 2.14 Comparison of Loadman and FWD at various test points (Whaley, 1994).
Figure 2.15 Correlation between Loadman and FWD (Whaley, 1994).
22
Despite the low correlation coefficient, it is noticed that both the FWD and the Loadman
follow the same overall trend. Whaley (1994) explains the poor correlation (low
correlation coefficient) to the fact that only the upper portion of the pavement section was
loaded with the Loadman, whereas the FWD loaded all layers resulting in a lower
modulus, representing the stiffness of all layers, not just the AC layer. Additionally, the
significant variations between the two devices may be caused by the presence of large
aggregate directly beneath the loading plate of the Loadman. These conclusions are
mirrored by Gros (1993).
Davies (1997) investigated similarities between the Loadman PFWD, Benkelman
Beam, and FWD on two thin membrane surface (TMS) roads. More detailed results of
this study are provided by Saskatchewan Highways & Transportation (SHT) (1998). The
TMS roads consisted of a 20 to 25 mm (0.79 to 0.98 in.) layer of oil treatment (cold mix)
for surfacing, with the application of a minimum of three to six sand seals since
construction. This was underlain by a graded/compacted subgrade (Davies, 1997). Each
test section consisted of 20 test points. The FWD was the first instrument over the test
point followed by the Benkelman Beam and the Loadman. Three seating drops and four
drops with each of four weights were used with the FWD. One measurement was taken
with the Benkelman Beam and ten measurements were taken with the Loadman. The
average of all ten Loadman measurements was plotted in addition to the first Loadman
measurement, the FWD measurement, and the Benkelman Beam measurement. The
Benkelman Beam produced the largest deflections at virtually all points tested followed
by the FWD and Loadman, all following similar trends, with few exceptions. The overall
results are shown in Figure 2.16.
23
Figure 2.16 Comparisons of Loadman, FWD, and Benkelman Beam (SHT, 1998).
There was a relatively good correlation between the Loadman PFWD and FWD with a
correlation coefficients of 0.86, significantly greater than obtained by Gros (1993) and
Whaley (1994). However, the PFWD and Benkelman Beam did not correlate as well,
yielding a correlation coefficient of 0.62. Additional results are shown in Table 2.1.
Table 2.1 Correlation coefficients of 20 test points on two TMS structures (SHT, 1998).
From the results shown, Davies (1997) reasoned that the PFWD did well to differentiate
between surfaces whose stiffness’ may be characterized as “soft” and “hard” but lacked
the ability to differentiate between surfaces that were of “hard” and medium stiffness.
Furthermore, Davies (1997) noted several other differences between devices that may
24
have contributed to the skewed results. The depth of the road structure that is measured
by each device is different. Thus, a device measuring only the upper portions of the road
structure will have a higher deflection than one measuring to a significantly greater depth,
paralleling the thoughts of Whaley (1994). Fleming and Rogers (1995) determined that
the zone “significantly stressed” was roughly equal to 1.5 to 2.0 times the diameter of the
loading plate. Also noted was the possibility that the Loadman is unable to deflect, thus
measure stiffness, above a certain level of stiffness. Davies (1997) suggests that this may
be the tradeoff of having an instrument small enough and light enough for one person to
use.
Livneh (1997) developed a double testing method to determine asphalt layer
moduli. Two measurements were necessary, one on the asphalt surface, and one at the
same point once the asphalt layer had been cored to its bottom. Utilizing this method, the
modulus of all the structural layers could be measured as well as the partial surface
modulus of all the layers not including the upper drilled asphaltic layer. The tested
section consisted of 120 to 250 mm (4.7 to 9.8 in.) thick asphalt layer underlain by
approximately 100 mm (3.9 in.) of variable size granular base and subbase layers. The
backcalculated FWD modulus values were larger than the modulus values determined
from the PFWD double testing, contradicting the results obtained by Gros (1993),
Whaley (1994), and Davies (1997). To compensate for differences in geometry, contact
pressure, and pulse loading between the two devices, correction factors were applied.
Original and corrected results are shown in Figure 2.17. Encouraging results were gained
only from the double PFWD testing (Livneh, 1997).
25
Figure 2.17 (a) Uncorrected and (b) corrected relationship between FWD modulus and PFWD modulus using the double testing technique (Livneh, 1997).
Livneh, et al. (1998) conducted tests aimed at developing correlations between the
deflection obtained with the Loadman PFWD and the central deflection obtained by
means of the Benkelman Beam. Side by side tests were completed at three different sites.
Site A consisted of 80 mm (3.1 in.) of asphalt underlain with 320 mm (12.6 in.) of
granular material. Site B consisted of 100 mm (3.9 in.) of asphalt and 450 mm (17.7 in.)
of granular material, and Site C had 160 mm (6.3 in.) of asphalt underlain by 1050 mm
(41.3 in.) of gravel. The results from each of the three sites are presented in Figure 2.18.
The results of the regression analysis indicate that the deflections obtained from the two
units correlated poorly with one another. The author attributed the differences to the fact
that the Loadman PFWD is only capable of measuring the properties of pavements of
limited thickness. The effect of deep layers on the surface deflection is a function of the
ratio between the layer thickness and the loading plate radius.
26
Figure 2.18 Loadman PFWD deflection versus Benkelman Beam deflection (Livneh, et al., 1998).
Honkanen (1991) performed a comparison of the Loadman and FWD on a gravel
road bound by oil. The results were consistent with other investigators. Loadman
produced higher modulus values than the FWD, with both instruments following a
similar trend, as seen in Figure 2.19.
Figure 2.19 Comparison of multiple FWD and PFWD measurements at one location (Honkanen, 1991).
27
2.2.3.2 Unbound Layers
Honkanen (1991) and Gros (1993) compared FWD, plate bearing tests, and
Loadman PFWD measurements on unbound surfaces. Honkanen (1991) completed
procedures on multiple test beds of equal length, differing thickness, and varying grain
size. Test results on sections with varying thicknesses revealed that as layer thickness
increases, modulus values from each instrument approach one another. Honkanen found
that to achieve sufficient reliability of the Loadman, it was necessary to perform multiple
measurements at each test point, a conclusion reiterated by Gros (1993) and Groenendijk,
et al. (2000). Honkanen (1993) indicated that at least four measurements should be
obtained at each test location. Of the four measurements, the first two should be
discarded and the remaining two should be averaged and used as a representative value.
Gros (1993) conducted tests on unbound aggregate containing sand, gravel, and
crushed gravel. Excluding errant test results, the correlation coefficients between the
Loadman and FWD ranged from 0.31 to 0.99 with an average of 0.77, which is higher
than for most of the studies with pavement layers. The Loadman and plate bearing test
yielded almost identical results. Modulus values given by the Loadman were at all points
larger than those given by the FWD and plate bearing unit. Results for unbound material
were shown to correlate well with one another, as shown in Figures 2.20 and 2.21. Both
the Loadman and the FWD follow the same trend with the FWD resulting in higher
modulus values on all points tested.
28
Figure 2.20 Comparison of Loadman and FWD at various test points (Gros, 1993).
Figure 2.21 Correlation between Loadman and FWD (Whaley, 1994).
Whaley (1994) also compared the results of testing the Loadman PFWD with the
FWD, Clegg Hammer, and Benkelman Beam to determine its efficiency when testing
deflection and layer moduli of base course layers. A comparison of all devices is shown
in Figure 2.22.
29
Figure 2.22 Comparison of FWD, Loadman PFWD, Benkelman Beam, and Clegg Hammer at various test points (Whaley, 1994).
Whaley (1994) concluded that the Loadman and FWD were best suited for this
application while the Benkelman Beam showed some discrepancies. The Clegg Hammer
showed no relation to the other devices. Comparing the FWD and Loadman, the
correlation is very near ideal, almost returning an error free 1 to 1 correlation (Whaley,
1994). The FWD moduli were slightly larger than those of the Loadman at 86% of points
tested, possibly due to the stress dependent nature of the aggregate.
Siekmeier, et al. (2000) performed comparative testing utilizing the PFWD for the
Minnesota Department of Transportation (Mn/DOT). Siekmeier, et al. (2000) completed
testing for 13 different sections utilizing the DCP, Loadman PFWD, SSG, and Dynatest
FWD. Test locations were primarily composed of a sand and gravel mixture with less
than 10 percent fines. Sand cone and nuclear density gauge density tests were conducted.
For the DCP testing, the penetration for each drop was recorded. With the PFWD, for
each test location, five tests were performed and recorded with the average of the last
30
three used for the modulus calculation. For the SSG, two tests were conducted for each
test location with the second being used for calculation purposes. If the two tests differed
by more than 3 percent, the tests were repeated at a new location. Results are shown in
Figure 2.23.
(ng = nuclear gage; sc = sand cone)
Figure 2.23 Moduli versus location for granular base material (Siekmeier, et al., 2000).
Locations 1 and 2 were beneath the inside wheel path, 3 and 4 between wheel paths, and
5 and 6 beneath the outside wheel path. The results indicated that all of the devices used
detected a variation in stiffness at each location, and all displayed a similar trend,
differing only in magnitude. Siekmeier, et al. (2000) explained the differences as a result
of the stress condition imposed by the instrument used. The lower the vertical stress
induced by the device, the lower the resulting modulus. Tests were also performed on a
mixture of clayey and silty sand fill. Like previous tests, each instrument showed a
similar trend. Siekmeier, et al. (2000) determined there was a strong correlation between
31
the instruments designed to measure modulus and that it is important to consider the
stress imposed by the instrument when stress dependent materials are used.
There was little agreement between percent compaction and modulus values.
Siekmeier, et al. (2000) explained that it is not realistic to know the Proctor maximum
density for every soil type found on a construction site. During the time of the testing
twelve Proctor tests had been completed covering typical soil types, however, did not
perfectly match each soil mixture that could be found at the location of an in situ density
test. The best available Proctor value was used, and as a result the agreement was poor.
It was concluded that compaction tests could be compared to in situ modulus tests only
when the material is uniform with respect to a single maximum Proctor density
(Siekmeier, et al., 2000).
Pidwerbesky (1997) completed a study utilizing the Clegg Hammer, FWD, NDM
and Loadman PFWD to determine their suitability for quantifying the present condition
and predicting the rutting potential of unbound granular base courses. The pavement
structure consisted of 90 mm (3.5 in.) of hot mix asphalt (HMA) over 200 mm (7.9 in.) of
crushed rock base course over a silty clay subgrade with a California Bearing Ratio
(CBR) of 12% (Pidwerbesky, 1997). The laboratory resilient modulus of the subgrade
and base course material was 105 MPa and 280 MPa, respectively. A Simulated Loading
and Vehicle Emulator (SLAVE) was used to load the pavement structure for
approximately one year. Once loading was complete, trenches were cut through the
layers for measuring density and modulus of the base course and subgrade. The
backcalculated moduli from FWD measurements were larger than the Loadman moduli at
86% of all points tested. Like the results given by Whaley (1994), the Clegg Hammer did
32
not correlate well with any of the other testing devices. The Loadman is not capable of
differentiating the moduli of various layers within a multi-layered pavement system, but
it can give an indication of the modulus of the layer being tested (Pidwerbesky, 1997). A
regression analysis of Loadman and FWD moduli was performed and shown in Figure
2.24.
Figure 2.24 Correlation between Loadman and FWD moduli (Pidwerbesky, 1997).
Pidwerbesky (1997) reported that the Loadman is substantially faster than NDM
methods, enabling increases in testing area and frequency. Further, Loadman is also
simpler to operate and interpret (a trained technician is not required), and does not have
to be calibrated for each material, which should be done with NDM tests.
Fleming, et al. (2000) performed FWD, German Dynamic Plate (GDP), TRL
Foundation Tester (TFT), and Prima 100 PFWD tests on two specially constructed trial
foundations. The GDP apparatus consists of a 10 kg (22 lb) falling mass impacting a
rubber buffer connected to a bearing plate. The drop height of the falling mass is set such
33
that the peak applied force is 7.07 kN (1589 lbf), i.e. 100 kPa (14.5 psi) stress. The
falling mass is guided by a vertical rod. The loading plate houses a velocity transducer
that measures the impact signal (Fleming, et al., 2000). The TFT consists of a 10 kg (22
lb) falling mass inside a guide tube that impacts a 300 mm (11.8 in.) load plate. The drop
height may be varied up to a maximum of 1 m (3.3 ft). The loading plate can be reduced
to 200 mm (7.9 in.) in diameter in order to vary the maximum contact stress, and to
provide a range of approximately 20 to 400 kPa (2.9 to 58 psi). A load cell and velocity
transducer is used to measure the force and surface deflection.
Typically, the sites tested consisted of a granular subbase at a depth of 37.5 mm
(1.5 in.) and/or crushed rock granular capping at a depth of 75 mm (3 in.), and a clayey
subgrade. Ten test locations were used at each site. The correlation coefficient and
coefficient of variation was determined for each test location (Fleming, et al., 2002).
Results are shown in Table 2.2 and 2.3.
Table 2.2 Summary of correlations between the FWD and GDP, TFT, and Prima 100 PFWD at Mountsorrel and Bardon test sites (Fleming, et al., 2000).
GDP TFT Prima 100 PFWD Subgrade Formation Details
Capping Test CC R2 CC R2 CC R2
Silty Clay subgrade 0.59 0.83 0.96 0.922 - - Gravelly
Silty Clay
150 mm subbase over up to 450 mm 6F1
capping capping 0.63 0.33 1.13 0.37 0.97 0.6
34
Table 2.3 Summary of correlations between the Prima 100 PFWD and GDP and TFT at Mountsorrel and Bardon test sites (Fleming, et al., 2000).
GDP TFT Prima 100 PFWD Subgrade Formation Details
Capping Test CC R2 CC R2 CC R2
Silty Clay subgrade - - - - - - Gravelly
Silty Clay
150 mm subbase over up to 450 mm 6F1 capping capping 0.63 0.38 1.13 0.53 - -
Fleming, et al. (2000) determined that the Prima 100 PFWD correlates reasonably with
the FWD, yielding a coefficient of variation of 0.6 at the same site where the TFT and
GDP yielded 0.37 and 0.33. A typical set of data for tests on 400 mm (15.7 in.) of
capping and the clayey subgrade is shown by Fleming, et al. (2002) in Figure 2.25.
Figure 2.25 Relationship between stiffness modulus determined by the portable dynamic plate test devices and the FWD (on subgrade and 400 mm thick granular capping) (Fleming, et al., 2000).
Thom and Fleming (2002) present a theoretical model that predicts the response
of a dynamic plate test. The dynamic model uses the Kθ non-linear model for unbound
35
materials (modulus = K1θK2, where θ is the sum of the principal stresses and K1 and K2
are constants). The second simplification is the use of a “load spread angle” technique
for stress analysis, which effectively turns the problem into a 1 dimensional one (Thom
and Fleming, 2002).
Four devices (FWD, Prima 100 PFWD, German Dynamic Plate, and TFT) were
theoretically applied to eight different foundations described in Table 2.4, and results
from the tests are shown in Table 2.5.
Table 2.4 Description of foundations applied to theoretical model (Thom and Fleming, 2002).
Table 2.5 Predicted surface moduli from different dynamic plate test devices (Thom and Fleming, 2002).
36
It may be noted that there are some variations in the results presented above, however, for
the most part, there is uniformity. Primarily, the TFT predicts the highest modulus, while
for most test sections; the FWD presents the lowest modulus.
Kamiura, et al. (2000) conducted a series of tests with two different kinds of
PFWD’s (Handy Type Falling Weight Deflectometer (HFWD), Prima 100 PFWD) and a
static bearing test on a sandy soil and subbase. He found that the modulus depended on
the strain of the surface beneath the loading plate. Kamiura, et al. (2000) reports that
modulus values would be relatively the same for each subgrade material if the strain
levels were in the range of 10-3 to 10-4 percent. Kamiura, et al. (2000) used a k value to
evaluate subgrade stiffness. The k value is the ratio of the stress induced by a 300 mm
(11.8 in.) loading plate to the deflection of the ground surface equal to 0.125 cm (0.05
in.). Comparison of k values for the Prima 100 and HFWD at multiple test locations is
shown in Figure 2.26. Figure 2.27 displays the comparison of deflection ratio at different
test locations for the Prima and HFWD.
Figure 2.26 Comparison of k value from Prima 100 and HFWD at multiple test locations (Kamiura, et al., 2000).
37
Figure 2.27 Comparison of the variation in deflection ratio with the number of drops at one location (Kamiura, et al., 2000).
Nazzal (2003) evaluated the potential use of the SSG, DCP, and Prima 100
PFWD as reliable means to measure the stiffness characteristics of highway materials for
possible application in the quality control quality assurance (QC/QA) procedures during
and after the construction of pavement layers and embankments. In addition to these
devices, FWD, PLT, and laboratory CBR tests were performed. Laboratory tests were
conducted to determine the zone influenced by the SSG and Prima 100 PFWD.
Regression analyses were performed to develop correlations between devices.
Field tests were performed at four separate locations. Conditions on U.S.
Highway 190 consisted of 200 mm (8 in.) of crushed limestone base on top of 200 mm (8
in.) of lime treated subgrade. Testing was also completed on four sections of Louisiana
State Highway 182. Measurements on existing subgrade material were completed in
Section 1. Section 2 consisted of 250 mm (10 in.) of cement treated base overlying 300
mm (12 in.), Section 3 consisted of 300 mm (12 in.) of cement treated subbase atop
38
subgrade, and Section 4 was made up of 300 mm (12 in.) of lime treated subbase atop
subgrade. Testing on U.S. Highway 61 was completed during the compaction of a 300
mm (12 in.) thick layer of untreated subbase. Finally, six sections were constructed at the
Louisiana Department of Transportation and Development (LA-DOTD) Accelerated
Load Facility (ALF). These sections included: one clayey silt soil, two cement stabilized
soils, one lime treated soil, one calcium sulfate hemihydrate (39.2% Calcium Oxide,
51.15% Sulfur Trioxide, 0.6% Silicon Dioxide, 0.75% Phosphorous Pent Oxide, 0.38%
Potassium, and 0.81% Aluminum Oxide), and one crushed limestone section (Nazzal,
2003). All layers were 300 mm (12 in.) in thickness and were constructed on existing
subgrade with the exception of the clayey silt material which was 100 mm (4 in.) thick.
Additionally, three trench sections (crushed limestone, sand, and recycled asphalt
pavement) were constructed at the ALF site to evaluate the devices for control of trench
backfill.
Test results from the ALF site indicated that the Prima 100 PFWD modulus
increased with increasing compactive effort as well as time after construction was
completed. This is shown in Figure 2.28. Similar moduli were obtained for the Prima
100 PFWD and the SSG. Results of the statistical analysis show that good correlations
do exist between the devices under evaluation (SSG, DCP, and Prima 100 PFWD) and
the standard tests (FWD, PLT, and CBR). Some of the results are shown in Figure 2.29.
39
Figure 2.28 Prima 100 PFWD measured increase in stiffness due to increased compactive effort and time (Nazzal, 2003).
Figure 2.29 Correlations between Prima 100 PFWD, FWD, and PLT (Nazzal, 2003).
40
All regression models had an adjusted R2 value and a significance level greater than 0.8
and 99.9% respectively (Nazzal, 2003). Laboratory results indicated that the depth
influenced by the SSG ranged from 180 to 190 mm (7.5 to 8.0 in.), while the Prima 100
PFWD influenced a depth of approximately 267 to 280 mm (10.5 to 11.0 in.). It was the
opinion of the author that the three devices in question could be reliably used to predict
the moduli obtained from PLT, FWD, and CBR values, and could be used to evaluate the
stiffness/strength parameters of different pavement layers and embankments.
2.3 PFWD AS TOOL TO EVALUATE THAW WEAKENING OF ROADS
Pavements in areas with seasonal freezing and thawing often undergo frost heave
and thaw weakening in addition to load-induced pavement distress. Vehicle traffic can
cause significant damage to roads that are weakened during the spring thaw. To
minimize damage, many road maintenance agencies impose load restrictions during
damage-susceptible periods. This is shown in Figure 2.30.
Spring thaw adversely affects pavement life while load restrictions impose local
economic hardships throughout the northern United States and Canada. Although the
maximum allowable load and the duration of the reduced load period vary widely among
agencies, they try to strike a balance between minimizing the disruption to the local
economy caused by the load restrictions and minimizing road damage.
41
Figure 2.30 Typical signage associated with placing load restrictions (Janoo and Cortez, 1998).
42
2.3.1 Current Methods to Evaluate Thaw Weakening of Roads
Pavement modulus is a key parameter in determining damage-susceptibility of
pavements. This can be monitored during spring thaw and through recovery using a
FWD. However, FWD purchase, operation, and maintenance is expensive. Second, even
if a state owns a FWD, it can only cover so many roads within a given time frame. As a
result, determining when the road has thawed and recovered sufficient strength to remove
the restriction is left to personal experience and subjective judgment.
Kestler, et al. (2000) distributed a survey to 45 state DOTs and multiple U.S.
Department of Agriculture Forest Service offices. The survey was aimed at determining
current load restriction practices. Of the 45 state DOTs that were solicited, 36 responded.
Three USFS regional offices also replied. Figure 2.31 shows a breakdown of methods
used for determining when to impose and when to remove load restrictions (Kestler, et
al., 2000). Kestler, et al. (2000) observed a majority of the respondents that post load
restrictions used subjective techniques, such as observation, to both place and remove
load restrictions. Additionally, many of the states indicated that they posted restrictions
only after the first signs of pavement distress are observed. Many also indicated that their
preference would be to switch from current subjective methods to more quantitative
methods (such as a FWD) if adequate resources were available (Kestler, et al., 2000).
Twenty four percent of responding DOTs were currently using quantitative methods to
place load restrictions and only 14% used the same methods to remove load restrictions.
The remaining 10%, who use quantitative methods to place load restrictions, simply keep
restrictions in place for a specific length of time or remove restrictions subjectively.
43
Roughly one quarter of the responding states use dates to impose restrictions. Some
roads were being restricted for as long as one third of one year. The length of seasonal
load restrictions for respondents is shown in Figure 2.32.
Figure 2.31 Methods for determining when to place and remove load restrictions (Kestler, et al., 2000).
44
Figure 2.32 Length of time over which load restrictions are placed (Kestler, et al., 2000).
According to most respondents, the longer the load restrictions are in place, the more
complaints were received from loggers and contractors. Table 2.6 lists typical DOT
responses regarding user feedback (Kestler, et al., 2000).
Table 2.6 Road user feedback to DOTs and USFS on spring thaw load (Kestler, et al., 2000).
45
The United States Federal Highway Administration (FHWA) investigated the
benefits of seasonal load restrictions in 1990. Table 2.7 shows the expected increase in
pavement life associated with varying pavement load restrictions. It is clear that seasonal
load restrictions can significantly extend the useful pavement life.
Table 2.7 Benefits from seasonal load restrictions (FHWA, 1990).
Pavement Load Reduction During Thaw
(%)
Expected Pavement Life Increase
(%) 20 62 30 78 40 88 50 95
Although seasonal weight restrictions extend pavement life, they also affect the
productivity of the trucking industry. In 1982, The Alaska Department of Transportation
and Public Facilities reported the statewide loss in revenue to the trucking industry was
approximately $100,000 USD per restricted day (1982 dollars). However, the associated
damage to state roads imposed when load restrictions were not enforced was roughly
$158,000 USD per day (C-SHRP, 2000).
The Canadian Strategic Highway Research Program (C-SHRP) prepared a
technical brief summarizing a series of presentations that were made to discuss policies
of seasonal pavement load restrictions in Canadian provinces, the United States and
various European countries. The methods used to determine when to place and remove
the weight restrictions are similar to those discussed by Kestler, et al. (2000). Direct
methods include the use of frost tubes or deflection testing, while indirect methods
include the use of historical databases, weather forecasts, prediction models or expert
46
judgment (C-SHRP, 2000). The report notes that each Canadian province utilizes weight
restrictions during spring thaw in an attempt to minimize damage. However, the
regulations vary not only in duration and extent but also on technical criteria and agency
practices. Most Canadian agencies impose spring load restrictions during March, and
remove the restrictions in May. Deflection testing is used by 7 of the 10 provinces, while
frost tubes are used primarily in British Columbia and Quebec (C-SHRP, 2000).
According to the brief, in the United States, 19 states have adopted the use of load
restrictions but there is no consistency between states in terms of where and when to use
the restrictions, how to apply them, and by how much to restrict the loads (C-SHRP,
2000). Finally, spring load restrictions are also used in France and several Scandinavian
countries. The amount of the load restrictions are presented, however, there is no
discussion of the methods used to develop them or methods used to determine when to
place and remove the restrictions.
Van Deusen (1998) reports on improved predictive equations for estimating thaw
duration based on deflection and environmental data collected from eight different low-
volume flexible pavement test sections in the state of Minnesota. Various studies have
found that air and subsurface temperatures can be used to identify thawing events. A
significant amount of research in this area has come from work done in the state of
Washington (Rutherford, et al., 1985; Mahoney, 1985; Rutherford, 1989), where it was
found that the onset of the critical period could be estimated with the air temperature
thawing index and that air temperature (air freezing index) can be used to predict the
duration of the thaw (Van Deusen, 1998). Results from the Minnesota test sites were
used to develop an equation to determine the duration of the thaw in terms of the freezing
47
index, and compare the results to those obtained by using the equation developed in
Washington. As a result, an improved thaw duration prediction relationship was
developed including the effects of frost depth. The predicted thaw durations from the
Washington equation appeared to be conservative in comparison to those determined
from the new Minnesota equation. It was recommended that the equation be validated
with one more winter/spring season of data (Van Deusen, 1998).
Research conducted by Rutherford, et al. (1985) in cooperation with the Federal
Highway Administration (FHWA) and the Washington State Department of
Transportation (WSDOT) aimed at providing procedures that would aid in determining
the amount of the restriction, where to apply them, and when to place and remove them.
The researchers investigated relationships between the Freezing Index (FI), Thaw Index
(TI), and thaw duration. Regression equations were developed from heat flow
simulations and are shown in Equations 2.4 and 2.5.
Eqn 2.4 )(018.025 FID +=
Where: D = thaw duration (days) FI = Freezing Index (°C days)
)(3.0 FITI = Eqn 2.5
Where: FI = Freezing Index (°C days)
The authors found that using the equations predicted the thaw duration to be longer than
the actual duration, however, WSDOT adopted the technique and used the equations to
predict the length of the thaw for Washington State. In addition, the researchers found
that as the amount of load reduction is increased there is an increase in pavement life.
48
Furthermore, thin or flexible pavements and unpaved roads require greater load reduction
during the spring thaw. The remainder of the findings are provided in Table 2.8.
Table 2.8 Summary of recommendations made by Kestler after Rutherford, et al. (1985).
Which pavements require load restrictions?
• Pavement with surface deflections 45-50% higher during spring thaw than summer.
• Pavements with frost susceptible base and subgrade materials. • Pavements with subgrade soils classified as ML, MH, CL, and CH. • Pavements where local experience so indicates. This includes poor ditch
drainage, high groundwater levels, etc. • Pavement in which distress has been observed (fatigue cracking and rutting).
When should vehicle load restrictions be placed? • When pavement surface deflections reach values 40-50% higher than summer
values. • When the air thaw index accumulates to values that correspond to thaw depths in frost susceptible materials. • When the thaw depth enters the frost susceptible subgrade materials as shown
by temperature measurements, frost tubes, or electric resistance gauges. What magnitude of load restrictions should be required?
• Allow only load levels that limit pavement deflections and strains to those estimated for summer conditions.
• Use load reductions that correlate with the desired increase in service life When should load restrictions be removed?
• When measured pavement surface deflections have returned to summer values (or design values).
The Minnesota Department of Transportation (Mn/DOT) conducted a study
aimed at evaluating the criteria used to predict when to place and remove load
restrictions. The approach was used to evaluate Minnesota’s load restriction practice, and
suggest improvements that would result in a more simple and accurate procedure. More
specifically, the objectives of the study were to (1) develop improved predictive
equations for estimating when to begin and end load restrictions, (2) investigate changes
49
in pavement strength in relation to freeze-thaw events, and (3) compare aggregate base
strength-recovery characteristics and assess their performance.
Eight test sections at the Mn/Road test facility were used for the study. Hot mix
asphalt (HMA) thicknesses range from 75 to 150 mm (3 to 6 in.). Six of the eight
sections are conventional designs with varying base and subbase thickness and two
sections are full-depth HMA sections. The base and subbase materials are dense-graded
sand and gravel mixtures of varying quality (Ovik, et al., 2000). In addition,
environmental data consisting of air temperature, frost depth, and subsurface temperature
were analyzed to determine the dates, on which the thaw began and ended, and to
determine the actual frost and thaw depth during spring-thaw. Finally, deflection data
from Mn/ROAD was used to determine the reduction in stiffness for the base and
subgrade materials during spring thaw (Ovik, et al., 2000).
Using the Freezing and Thaw Index data as well as the observed thaw duration,
the prediction equations developed by WSDOT were adjusted to more accurately predict
the thaw duration for conditions in Minnesota. This is shown below. Using Equations
2.4 and 2.5 directly proved that they predict thaw too late in Minnesota.
FIPPFID ⋅−++= 120901.19010.015.0 Eqn. 2.6
Where: D = thaw duration (days) P = frost depth (m) FI = Freezing Index (°C days)
In addition, historical posting dates from 1986 through 1998 were compared to the
posting dates predicted using the new technique. It was found that there was typically a
week or more delay from the time that load restrictions should be placed until restrictions
were actually posted (Ovik, et al., 2000).
50
The final result of the study was the adoption of a new procedure for placing load
restrictions in Minnesota. The policy uses actual and forecasted average daily
temperatures to determine when the restrictions should be placed (Ovik, et al., 2000).
2.3.2 Past Test Programs, Results, and Recommendations
Davies (1997) investigated the ability of the Loadman PFWD to track strength
changes through the spring thaw for Saskatchewan Highways & Transportation (SHT).
Five thin membrane surfaces (TMS) of varying age, construction history, annual
maintenance costs, and traffic volumes were tested. Each surface contained between
three and six test points, totaling 21 points. SHT currently utilizes the Benkelman Beam,
which is the standard against which the Loadman was measured. For each test point, five
Loadman measurements were taken and averaged for comparison to the Benkelman
Beam. Peak frost free strengths were established in late fall and testing continued once
the spring thaw began. After an evaluation of the initial regression analysis, it was
determined that temperature correction was required to compare the Loadman data
collected through the different seasons. As a result, data points were separated into three
groups based on their regression coefficients, deflection values observed at 20°C, and
DCP/coring information and field observations. The categories are as follows:
1. seal-on-subgrade
2. Thick, (relatively) soft mat; and
3. Thick, (relatively) hard mat.
A temperature correction of 10 °C was applied to all of the field values by means
of best fit equations for the respective category. Davies (1997) found this necessary in
51
order to compare the Loadman results through the seasons. Results are shown below in
Figure 2.33.
Figure 2.33 Comparison in the ability of the Loadman PFWD and Benkelman Beam to track strength change through spring thaw (Davies, 1997).
Davies (1997) noted “significant differences” in the ability of the PFWD to track strength
change, citing the varying layer thicknesses at the test locations as the prime reason.
Fleming and Rogers (1995) determined that the zone “significantly stressed” was roughly
equal to 1.5 to 2.0 times the diameter of the loading plate. From this, he concluded that
the Loadman follows the change in strength if the asphalt thickness is less than one half
(<100 mm) of the Loadman’s zone of influence.
2.4 PFWD QUESTIONNAIRE
In February 2003 and May 2004 a questionnaire, aimed at determining current
usage of the PFWD as an alternative to traditional compaction control devices and as a
tool to evaluate thaw weakening of roads, was distributed to each of the 50 state
Departments of Transportation (DOTs). The following sections describe the results of
the survey.
52
2.4.1 2003 Results
The survey distributed in March, 2003 had a response rate of 56%. Of the 28
states that responded, none had past or present experience in using the PFWD. However,
four of the respondents, or ~15%, were aware of current usage by other organizations.
2.4.2 2004 Results
The survey distributed in May, 2004 had a response rate of 36%. Of the 18 states
that responded, one (Minnesota) has past or present experience in using the PFWD.
2.5 SUMMARY
The PFWD compared marginally with other devices when testing on pavement
layers. Correlation coefficients relating the PFWD to other devices were relatively low.
The portable devices generally result in higher modulus values than the FWD, possibly
due to the thickness influenced. Several investigators have reported that the zone of
influence for the different models lies primarily between one and two loading plate
diameters. Large aggregate particles beneath the loading plate of the portable devices
also have been shown to affect the results, as the particles increase the resulting modulus
values.
The PFWD compared reasonably with other devices when testing on unbound
layers. The PFWD reported higher modulus values than the FWD and plate bearing unit
when used on granular soils. The range of correlation coefficients was similar to those
obtained on pavement layers. The differences between Loadman and the FWD for
53
measurements obtained on a bound surface were reversed on an unbound surface;
however, correlations between devices were much closer.
Few researchers have examined the methods used to evaluate when to place and
remove load restrictions. Kestler, et al. (2000) distributed a survey aimed at determining
current load restriction practices. A majority of respondents use subjective techniques to
aid in determining when load restrictions should be applied and removed. The lengths of
the restrictions ranged from as little as three weeks to, according to some respondents,
greater than 12 weeks. The Canadian Strategic Highway Research Program also
summarized its findings from a survey similar to the one distributed by Kestler, et al.
(2000). The findings were similar. The Washington and Minnesota Departments of
Transportation developed equations using the freezing and thawing indices to determine
the duration of spring thaw period and have had varying degrees of success putting them
into practice.
Davies (1997) performed testing aimed at determining the ability of the PFWD to
track seasonal stiffness variations. He concluded that the PFWD did adequately follow
strength change through spring thaw. This, however, is only valid if the asphalt thickness
is less than one half of the zone of influence of the PFWD.
54
(BLANK PAGE)
55
CHAPTER 3
FIELD & LABORATORY TEST PROTOCOL
3.1 INTRODUCTION
This test protocol presents the procedures for the field and laboratory evaluation
of the Portable Falling Weight Deflectometer (PFWD). This protocol includes the
following:
1. Field test site locations.
2. Current road condition, subsurface conditions, and cross sections of field test
sites.
3. Instrumentation description, and installation and monitoring procedures.
4. Descriptions of the field testing procedures and data gathered for the study of
seasonally posted low volume paved and unpaved roads.
5. Descriptions of testing procedures, materials tested, and data gathered for the field
and laboratory study of the compaction of subgrades and construction materials.
3.2 FIELD TEST SITE LOCATIONS
3.2.1 Seasonally Posted Low Volume Roads
The performance of seven paved and three gravel surfaced roads were monitored
during the spring of 2004. This portion of the project made use of existing instrumented
test sites. Three of the instrumented test sites were part of previous or ongoing New
England Transportation Consortium (NETC) and Maine Department of Transportation
(MaineDOT) projects constructed under the direction of Dr. Dana N. Humphrey. One of
56
the instrumented sites was part of ongoing research by the United States Forest Service
(USFS) under the direction of Maureen Kestler. Additional sites were selected in
consultation with NETC, USFS, MaineDOT, and the Vermont Agency of Transportation
(VAOT). A summary of each site is provided in Table 3.1 with approximate geographic
locations shown in Figure 3.1. Details are discussed in the following subsections.
3.2.1.1 Kennebec Road – Hampden/Dixmont, Maine
Kennebec Road is a seasonally posted, low volume, hot mix asphalt (HMA)
surfaced road. The road is owned and maintained by the State of Maine. This road was
selected because it has a relatively low traffic volume, is posted for seasonal weight
restrictions, and is in close proximity to the University of Maine campus. The conditions
at the site consist of approximately 152 mm (6 in.) of HMA pavement, and approximately
203 mm (8 in.) of subbase material. Typical road condition is shown in Figure 3.1.
Figure 3.1 Typical condition of Kennebec Road, Hampden/Dixmont, Maine in April, 2003.
57
Table 3.1 Summary of seasonally posted low volume road field test sites.
Project Name Location HMA
Thicknessmm (in.)
Subbase Aggregate Thickness mm (in.)
Subgrade Type
Instrumentation & Field
Measurements
Kennebec Rd Hampden / Dixmont, Me. 152 (6) 203 (8) Silty Sand
Thermocouples, Piezometers,
FWD
Lakeside Landing Rd Glenburn, Me. NA 203 (8) Sandy Silt
Thermocouples, Piezometers,
FWD
Stinson Lake Rd Rumney, N.H. 127 (5) 305-381 (12-15)
Silty Clay To
Silty Sand
Thermistors, Waterwell, TDR,
FWD
Buffalo Rd Rumney, N.H. 127 (5) 300 (12) Silty Sand Thermocouple, Frost Tube, FWD
USFS Parking Lot Rumney, N.H. NA 180 (7) Sandy Silt Thermistors, TDR,
Waterwells, FWD Crosstown Rd Berlin, Vt. NA 300 (12) Silty Sand Thermocouple, FWD Knapp Airport
Parking Lot Berlin, Vt. 127 (5) 300 (12) Silty Sand Frost Tube, FWD
127 (5) 483 (19) 288 (11.3) Silty Clay
Thermocouples, Piezometers,
FWD
127 (5) 483 (19) 326 (12.8) Silty Clay
Thermocouples, Piezometers,
FWD
Witter Farm Road Orono, Me.
127 (5) 635 (25) Silty Clay Thermocouples,
Piezometers, FWD
150 (6) 600 (24) Silty Clay
to Silty Sand
Thermocouples, Piezometers,
FWD
150 (6) 300 (12) Silty Clay
to Silty Sand
Thermocouples, Piezometers,
FWD Route 126 Monmouth /
Litchfield, Me.
150 (6) 150 (6) grindings
Silty Clay to
Silty Sand
Thermocouples, Piezometers,
FWD
180 (7) 640 (25)
Silty Clay to
Sandy Gravel
Thermocouples, Piezometers,
FWD
180 (7) 640 (25)
Silty Clay to
Sandy Gravel
Thermocouples, Piezometers,
FWD Route 1A
Frankfort / Winterport, Me.
180 (7) 640 (25)
Silty Clay to
Sandy Gravel
Thermocouples, Piezometers,
FWD
58
Figure 3.2 Approximate geographic location of spring thaw test sites.
Two separate sections were chosen for testing. Each section was intrumented
with one thermocouple and two standpipe piezometers on November 14, 2003.
Groundwater was not encountered during installation of instruments in Section 1. The
groundwater table at Section 2 at the time of instrument installation was approximately
1.5 to 1.8 m (5 to 6 ft) beneath top of pavement. Results of laboratory tests on field
samples obtained during instrumentation are provided in Table 3.2. Details of the
instruments, as well as, installation and monitoring methods are described in subsequent
sections.
59
Table 3.2 Laboratory properties of in-situ material at Kennebec Road, Hampden/Dixmont, Maine.
Section 1 Section 2
Boring No.
Depth (m)
Water Content
(%)
Visual Classification
BoringNo.
Depth (m)
Water Content
(%)
Visual Classification
PZ 1 0.2 – 0.4 4.1 PZ 3 0.2 – 0.4 8.5
PZ 1 0.4 – 0.8 7.8 fine to medium
gravel PZ 3 0.4 – 0.5 11.0
TH 1 0.2 – 0.4 3.7 PZ 3 0.5 – 0.8 11.7
medium to course sand with a trace
of silt
TH 1 0.4 - 0.8 10.1 fine to medium
gravel TH 2 0.2 – 0.3 8.5
TH 1 0.8 – 1.4 11.9 TH 2 0.3 – 0.5 10.8
medium to course sand with a trace
of siltTH 1 1.4 – 1.6 10.7 TH 2 1.1 – 1.2 13.6
TH 1 2.0 – 2.3 18.1 TH 2 1.5 – 1.8 23.7
TH 1 2.3 – 2.6 14.8
fine to medium sand with a trace
of silt TH 2 1.8 – 2.1 18.3
PZ 2 0.2 – 0.4 4.4 TH 2 2.1 – 2.4 20.6
medium sand with a trace of silt
PZ 2 0.4 – 0.8 5.3 fine to medium
gravel PZ 4 0.2 – 0.5 7.6 medium to course sand with a trace
of silt
3.2.1.2 Lakeside Landing Road – Glenburn, Maine
Lakeside Landing Road is a seasonally posted, low volume, gravel surfaced road.
The road is owned and maintained by the Town of Glenburn. This road was selected
because it has a low traffic volume, is posted for seasonal weight restrictions, is in close
proximity to the University of Maine campus, and conditions significantly deteriorate
during spring thaw as shown in Figure 3.3. The road section consists of approximately
0.2 m (8 in.) of gravel overlying a geotextile placed on subgrade.
60
Figure 3.3 Typical road condition of Lakeside Landing Road, Glenburn, Maine in April, 2003.
Two separate sections of the road were chosen for testing. Each section was
intrumented with one thermocouple and two standpipe piezometers on November, 7,
2003. Groundwater was not encountered during installation. Results of laboratory tests
on samples obtained during instrumentation are provided in Table 3.3. Details of the
instruments, as well as, installation and monitoring methods are described in subsequent
sections.
Table 3.3 Laboratory properties of in-situ material at Lakeside Landing Road, Glenburn, Maine.
Boring No.
Depth (m)
Water Content
(%)
Visual Classification
TH 1 0.2 – 0.8 7.4 TH 2 0.2 – 0.8 8.9
fine to medium sandy gravel with a trace of silt
61
3.2.1.3 Stinson Lake Road– Rumney, New Hampshire
Stinson Lake Road is a seasonally posted, low volume, HMA surfaced road. The
road is owned by the Town of Rumney and is maintained by the State of New
Hampshire. The site was selected because of its close proximity to an existing test site at
the USFS Parking Lot. The conditions at the site consist of approximately 127 mm (5
in.) of HMA pavement, and approximately 305 to 381 mm (12 to 15 in.) of subbase
material.
Figure 3.4 Typical road condition of Stinson Lake Road, Rumney, New Hampshire in July, 2003.
Two thermistor probes, five time domain reflectometry (TDR) probes, and one
standpipe piezometer were installed on July 24, 2003. The groundwater table was
encountered during instrumentation at approximately 0.9 m (3 ft) beneath top of
pavement. Results of laboratory tests on field samples obtained during instrumentation
62
are shown in the Table 3.4. Details of the instruments, as well as, installation and
monitoring procedures are provided in a later section.
Table 3.4 Laboratory properties of in situ subbase material at Stinson Lake Road, Rumney, New Hampshire.
Boring No.
Depth (m)
Water Content
(%)
Visual Classification
0.0 - 0.5 4.2 0.5 - 0.9 6.8
fine to medium graveland some sand
0.9 - 1.4 14.3 1.4 - 1.8 14.0
T2
1.8 - 2.2 15.7
fine sandy gravel witha trace of silt
3.2.1.4 Buffalo Road – Rumney, New Hampshire
Buffalo Road is a seasonally posted, low volume, HMA surfaced road. The road
is owned and maintained by the Town of Rumney. Like Stinson Lake Road, Buffalo
Road was chosen for this project due to its closeness to the USFS Parking Lot site. The
conditions at the site consist of approximately 127 mm (5 in.) of HMA pavement,
overlying a silty sand subgrade. One thermocouple and one frost tube were installed in
December 2004. Results of laboratory tests on field samples obtained during
instrumentation are provided in Table 3.5. Details of the installation are described
elsewhere.
63
Table 3.5 Laboratory properties of in situ material at Buffalo Road, Rumney, New Hampshire.
BoringNo.
Depth (m)
Water Content
(%)
Visual Classification
0.0 – 0.3 11.7 0.3 – 0.9 15.3 0.9 – 1.2 21.2 1.2 – 1.5 19.4 1.5 – 1.8 24.3 1.8 – 2.1 27.1 2.1 – 2.4 28.6
TH 1
2.4 – 2.7 26.3
fine to medium silty sand
3.2.1.5 USFS Parking Lot – Rumney, New Hampshire
The gravel surfaced parking lot, located at the Rumney Rocks recreational area, is
owned and maintained by the USFS. The conditions at the site consist of approximately
178 mm (7 in.) of gravel overlying a sandy silt subgrade. The subgrade material is
classified as SP-SM according to the USCS. Existing site conditions are shown in
Figure 3.5. One thermistor probe, two TDR probes, and two standpipe piezometers were
installed during the winter of 2003 by the USFS. A gradation of the subbase material is
provided in the Appendix A.
64
Figure 3.5 Existing site conditions at USFS Parking Lot, Rumney, New Hampshire.
3.2.1.6 Crosstown Road – Berlin, Vermont
Crosstown Road is a seasonally posted gravel surfaced road. It is owned and
maintained by the Town of Berlin. This site was selected in consultation with VAOT due
to its close proximity to their offices and due to the extreme deterioration the road
experiences during spring thaw. Existing conditions are shown in Figure 3.6. The road
conditions consist of 305 mm (12 in.) of gravel overlying subgrade.
65
Figure 3.6 Existing conditions at Crosstown Road, Berlin, Vermont in March 2004.
One thermocouple and one frost tube were installed in February 2004.
Groundwater was not encountered during instrumentation. Results of laboratory tests on
field samples obtained during instrumentation are shown in Table 3.6. Installation details
are described elsewhere.
Table 3.6 Laboratory properties of in situ material at Crosstown Road, Berlin, Vermont.
Boring No.
Depth (m)
Water Content
(%)
Visual Classification
0.0 – 0.6 10.2 0.6 – 1.2 10.4
fine to medium sandy gravel
1.2 – 1.8 14.5 1.8 – 2.4 27.8
TH 1
2.4 – 2.7+ 25.4
medium to course sand and some silt
66
3.2.1.7 Knapp Airport Parking Lot – Berlin, Vermont
This site was selected because the VAOT utilizes space at the airport four housing
their FWD. In addition, it is in close proximity to the Crosstown Road test site. The
parking lot consists of 127 mm (5 in.) of HMA pavement and approximately 300 mm (12
in.) of subbase material overlying subgrade soil. One frost tube was installed in February
2004. No groundwater was encountered during installation. Laboratory test results on
samples obtained during installation are provided in the Table 3.7. Instrumentation
details are described in subsequent sections.
Table 3.7 Laboratory properties of in situ material at Knapp Airport Parking Lot, Berlin, Vermont.
Boring No.
Depth (m)
Water Content
(%)
Visual Classification
0.0 – 0.6 8.4 0.4 – 1.2 14.9 1.2 – 1.8 14.6
F1
1.8 – 2.4 13.97
medium to course sand and some silt
3.2.1.8 Witter Farm Road – Orono, Maine
Reconstruction of a 77 m (255 ft) section of Witter Farm Road in the town of
Orono was completed in 1997 as part of NETC Project No. 95-1. The full scale field trial
was constructed to investigate the use of tire chip/soil mixtures to reduce frost penetration
and improve drainage of paved roads (Lawrence, et al., 2000). The project was divided
into six 12.2 m (40 ft) long paved sections. Three of the sections (Section 1, Section 2,
and Control Section) were used for this study. A plan view is shown in Figure 3.7.
67
Figure 3.7 Plan view of test sections at Witter Farm Road (Lawrence, et al., 2000).
Each of the three sections contains 127 mm (5 in.) of HMA bituminous pavement.
Section 1 contains 483 mm (19 in.) of gravel subbase (MaineDOT Type D) overlying 326
mm (12.8 in.) of a 33% / 67% tire chip/gravel mixture. Section 2 contains 483 mm (19
in.) of gravel subbase (MaineDOT Type D) overlying 288 mm (11.3 in.) of 67% / 33%
tire chip/gravel mixture. The Control Section contains 635 mm (25 in.) of gravel subbase
(MaineDOT Type D) and has no tire chips. Cross sections for each test section are
shown in Figure 3.8.
68
Figure 3.8 Cross section of test sections at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000).
The granular subbase met MaineDOT Specification 703.06, Type D (152 mm (6
in.) maximum size, 25 to 70% passing the 6.4 mm (¼ in.), 30% maximum passing the
No. 40, and 7% maximum fines). This material was used for subbase over the tire chips,
tire chip/soil mixtures, and for the subbase course in the Control Section (Lawrence, et
al., 2000). Gradation curves for this material as well as the tire chip/soil mixtures are
provided in Appendix A. The subgrade soil was classified as CL according to the USCS
and A-4(8) according the AASHTO classification system. Laboratory test results on
subgrade samples are shown in the Table 3.8. Gradations of subbase and subgrade
samples can be found in Appendix A.
69
Table 3.8 Laboratory index properties of cohesive subgrade material at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000)
Sample Type Location Depth
(m)
Water Content
(%)
PlasticLimit
LiquidLimit
Plasticity Index
SpecificGravity
F1 0.6 23.8 F2 0.6 21.1 F2 1.2 21.7 2.72
Boring
F2 1.8 22.0 2.74 Section 1 20.4 Section 2 24.1 24 32 8 2.64 Bulk
Subgrade Control 17.7 21 27 6 Section 1 0.0 – 0.9 20.4 Section 1 0.9 – 1.5 15.5 Section 2 0.0 – 0.9 20.7 Section 2 0.9 – 1.5 14.4 Control 0.0 – 0.9 20.7
Auger
Control 0.9 – 1.5 32.1
Each test section contains one thermocouple string. Three standpipe piezometers
were installed at locations P1, P2, and P3 shown in Figure 3.7. Instrumentation details
are discussed in the following sections.
3.2.1.9 Route 126 – Monmouth/Litchfield, Maine
Reconstruction of a 9.5 km (5.9 mile) section of Route 9 in the towns of
Monmouth and Litchfield, Maine was completed in 2002 as part of NETC Project No.
00-8. The purpose of the full scale field trial was to investigate the effectiveness of
geogrid and drainage geocomposite with thin pavement sections and soft subgrade soils
in cold regions (Helstrom and Humphrey, 2005). The project was divided into 12
sections. Sections 1 through 5 are 60 m (196.8 ft) long, Sections 6 through 10 are 40 m
70
(131.2 ft) long, and Sections 11 and 12 are 20 m (65.6 ft) long. Three control sections
(Section 3, Section 8, and Section 12) were used for this study.
Each control section contains 150 mm (6 in.) of HMA pavement. Sections 3 and
8 contain 300 mm (12 in.) and 600 mm (24 in.) of subbase (MaineDOT Type D
aggregate), respectively. Section 12 serves as one of two reclaim sections for this
project. This section contains 76 to 152 mm (3 to 6 in.) of reclaimed asphalt grindings
overlying existing subbase material. Cross sections of all sections are shown in Figure
3.9.
The granular subbase met MaineDOT Specification 703.06, Type D (152 mm (6
in.) maximum size, 25 to 70% passing the 6.4 mm (¼ in.), 30% maximum passing the
No. 40, and 7% maximum fines). It is classified as A-1-a according to the AASHTO
classification system. Gradation curves for this material are provided in Appendix A.
A subsurface investigation reported very poor subgrade soils throughout the
length of the project (Fogg, 2002). These soils are moist and plastic with standard
penetration blow counts as low as 10. The water content, liquid limit, plasticity index,
and classification of the subgrade samples taken during the investigation are summarized
in Table 3.9. Grain size distributions of the subgrade samples are given in Appendix A.
71
ASPHALT
SUBBASE
SUBGRADE
REINFORCEMENTGEOGRID
150
300
150
SUBGRADE
REINFORCEMENTGEOGRIDSUBBASE
ASPHALT
150
300
ASPHALT
SUBBASE
SUBGRADE
150
SUBGRADE
DRAINAGEGEOCOMPOSITE
SUBBASE
ASPHALT
150
300
ASPHALT
SUBBASEDRAINAGEGEOCOMPOSITE
SUBGRADE
150
SUBGRADE
REINFORCEMENTGEOGRID
SUBBASE
ASPHALT
600
SUBGRADE
REINFORCEMENTGEOGRIDSUBBASE
150 ASPHALT
SUBGRADE
ASPHALT150
SUBBASE600
SUBGRADE
DRAINAGEGEOCOMPOSITE
600 SUBBASE
150 ASPHALT
DRAINAGEGEOCOMPOSITE
SUBGRADE
ASPHALT150
SUBBASE
150 ASPHALT
EXISTING BASE
PAVEMENT GRINDINGS75-150DRAINAGEGEOCOMPOSITE
EXISTING BASE
75-150 PAVEMENT GRINDINGS
ASPHALT150
TEST SECTION 11+520 - 1+580
150
150
TEST SECTION 21+580 - 1+640
TEST SECTION 31+640 - 1+700
TEST SECTION 41+700 - 1+760
TEST SECTION 51+760 - 1+820
REINFORCEMENTGEOGRID
150
150
TEST SECTION 63+940 - 3+980
300
300
TEST SECTION 73+980 - 4+020
TEST SECTION 84+020 - 4+060
TEST SECTION 94+060 - 4+100
REINFORCEMENTGEOGRID300
300
TEST SECTION 104+100 - 4+140
TEST SECTION 114+680 - 4+700
TEST SECTION 124+700 - 4+720
*All dimensions shown are in millimeters
Figure 3.9 Test section layout of Route 126 Monmouth/Litchfield, Maine (Helstrom and Humphrey, 2005).
72
Table 3.9 Laboratory index properties of subgrade material at Route 126, Monmouth/Litchfield, Maine (Helstrom and Humphrey, 2005).
Classification Boring &
Sample ID Test
Section Station
(m) Offset
(m) Depth
(m) WC LL PI USCS AASHTO
HB-MONM
103/TS3-1 3 1+670 RT Subgrade 14.0 -- -- CL-ML A-4
HB-MONM
103/TS3-2 3 1+670 RT Subgrade 21.1 27 10 CL A-4
8-3 8 4+040 0.9RT 1.5-1.8 23.3 N P CL-ML A-4
8-4 8 4+040 0.9RT 1.8-3.0 27.3 24 6 CL-ML A-4
12-2 12 4+710 0.9RT 2.4-3.0 -- -- -- -- --
Each test section contains one thermocouple string and two vibrating wire
piezometers, details of which are presented elsewhere.
3.2.1.10 Route 1A – Frankfort/Winterport, Maine
Reconstruction of a 3.06 km (1.9 mile) section of Route 1A in the towns of
Winterport and Frankfort was completed in 1997 as part of a MaineDOT research project.
Reinforcement geogrid, reinforcement geotextile, separation geotextile, and high
compressive strength geocomposite drainage net were used in this project to evaluate
their reinforcement, separation, filtration, and drainage performance for Maine soil and
climatic conditions (Fetten and Humphrey, 1998). The project was divided into six
sections, each with different geosynthetic applications. Three of the sections (D-1, D-2,
and E-3) were used for this study.
73
The test sections considered for this study have 180 mm (7 in.) of HMA pavement
and 640 mm (25 in.) of aggregate subbase. Section D-1 has a drainage geocomposite
located 460 mm (18 in.) beneath the subbase / subgrade interface. Section D-2 has a
drainage geocomposite on subgrade, and Section E-3 has one layer of reinforcement
geogrid 250 mm (9.8 in.) beneath the asphalt / subbase interface. Cross sections of the
sections are shown in Figure 3.10.
The subbase aggregate used for this project was uniformly graded sandy gravel.
Approximately 50% of the soil is between 12 mm and 75 mm (0.47 in. and 3.0 in.). It is
classified as an A-1-a soil according to the AASHTO classification system. The
gradation of this material is shown in Appendix A.
A subsurface investigation (Hayden, 1996) reported poor subgrade soil conditions
along the entire length of the project. Moist clay soils were the dominant soil type
encountered. The material is plastic with water contents greater than 20% in some areas
and liquid limits as high as 37 and plasticity indexes of 17. Natural water contents for
samples taken at thermocouple locations are provided in Table 3.10. The samples were
classified as A-6 according to the AASHTO classification system. Three laboratory CBR
tests were conducted on representative samples at Modified Proctor optimum moisture
contents. These tests produced values of 2.6, 3.2, and 3.6 (Fetten and Humphrey, 1998).
74
Figure 3.10 Test section layout of Route 1A Frankfort/Winterport, Maine (Fetten and
Humphrey, 1998).
75
Table 3.10 Water contents at thermocouple location on Route 1A Frankfort/Winterport, Maine (Fetten and Humphrey, 1998).
Section Station (m)
Offset(m)
Depth (m)
Water Content
(%) 1.04 7.5 1.65 7.7 E-3 88+85 NA 2.26 7.0 0.2 21.5 CL 0.6 23.1 0.2 21.7
0.2 – 0.6 21.7 2R 0.6 – 1.1 22.7
0.2 21.0
D-1 78+00
2L 0.2 – 1.1 23.3 NA – information not available.
Section D-2 contains two vibrating wire piezometers. The U.S. Army Corps of
Engineers Cold Regions Research and Engineering Laboratory (CRREL) also installed
two thermocouple strings.
3.2.1.11 Route 11, Wallagrass Plantation, Maine and Route 167, Presque Isle/Fort Fairfield, Maine.
Routes 11 and 167 were used for testing on one day during the spring of 2003 as
part of an ongoing MaineDOT research project (Bouchedid and Humphrey, 2004). Four
test sections were used at each site. A detailed description of each test section is
provided in Table 3.11 and 3.12.
76
Table 3.11 Test Section description of Route 11, Wallagrass Plantation, Maine (Bouchedid and Humphrey, 2004).
Location Asphalt Thickness mm (in.)
Subbase Thicknessmm (in.)
Underlying Material
Test Pit 1 118 (4.6) 647 (25.5) Fill Test Pit 2 103 (4.1) 797 (31.4) Fill Test Pit 3 128 (5) 803 (31.6) Subgrade Test Pit 4 136 (5.4) 617 (24.3) Ledge
Table 3.12 Test Section description of Route 167, Presque Isle/Fort Fairfield, Maine
(Bouchedid and Humphrey, 2004).
Location Asphalt Thickness mm (in.)
Subbase Thicknessmm (in.)
Underlying Material
Test Pit 1 135 (5.3) 730 (28.7) Subgrade Test Pit 2 120 (4.7) 773 (30.4) Ledge Test Pit 3 118 (4.6) 895 (35.2) Ledge Test Pit 4 127 (5) 750 (29.5) subgrade
3.2.2 Compaction Control Field Test Sites The field sites for evaluation of subgrades and construction materials were located
with the assistance of the NETC Technical Committee assigned to this project and
MaineDOT. Field sites are summarized in Table 3.13, and each is discussed separately in
the following subsections.
77
Table 3.13 Summary of compaction control field test sites.
Project Name Location Material Types
Tested Measurements
I-84 Reconstruction Southington, Ct. Crushed Gravel, Subgrade PFWD, NDM
Rt. 25 Realignment Effingham/Freedom, N.H.
Construction Sand, Gravel PFWD, NDM
Rt. 26 Bypass New Gloucester, Me. MaineDOT Type D & E PFWD, NDM
Rt. 201 Reconstruction The Forks, Me. Reclaimed
Asphalt FWD, PFWD,
NDM Commercial Paving &
Recycling Scarborough, Me. Subgrade, Flexpave PFWD, NDM
3.2.2.1 I-84 Reconstruction – Southington, Connecticut
Reconstruction of several sections of I-84 in Southington, Connecticut took place
in the fall of 2003. One section was used for this project. Conditions consisted of
variable depths of crushed gravel overlying existing subgrade. Thicknesses ranged from
102 to 356 mm (4 to 14 in.), some of which had been subjected to varying degrees of
compaction. The test section is shown in Figure 3.11.
Figure 3.11 I-84 test section, Southington, Connecticut.
78
The crushed gravel was classified as A-1-a by the AASHTO classification system.
The crushed gravel has a maximum dry density of 2.31 Mg/m3 (144 lb/ft3) at an optimum
water content of 7.4% as determined by AASHTO T 180. Gradations of both materials
are provided in Appendix A.
3.2.2.2 Route 25 Realignment – Effingham/Freedom, New Hampshire
The realignment of a portion of Route 25 in the towns of Effingham and Freedom,
New Hampshire, took place in conjunction with a bridge replacement project. Two
sections were used for testing. Section 1, shown in Figure 3.12, was used for testing
crushed gravel meeting NHDOT Specification 304.3 (75 mm (3 in.) maximum size, 95-
100% passing the 50 mm (2 in.) sieve, 55-85% passing the 25 mm (1 in.) sieve, 27-52%
passing the No. 40 sieve, and 12% maximum fines). Section 1 consisted of 203 mm (8
in.) of crushed gravel overlying 203 mm (8 in.) of material meeting NHDOT
Specification 304.1 (70-100% passing the No. 40 sieve, and 12% maximum fines).
Section 2, shown in Figure 3.13, consisted of 305 mm (12 in.) of sand overlying
subgrade. Both materials were classified as SW according to the USCS and A-1-a
according to AASHTO. The crushed aggregate has a maximum dry density of 1.92
Mg/m3 (120 lb/ft3) at an optimum moisture content of approximately 12%. The sand
aggregate has a maximum dry density of 2.17 Mg/m3 (135 lb/ft3) at an optimum moisture
content of 11%. Gradation and moisture density curves for each material are shown in
Appendixes A and B, respectively.
79
Figure 3.12 Route 25 Test Section 1, Effingham/Freedom, New Hampshire.
Figure 3.13 Route 25 Test Section 2, Effingham/Freedom, New Hampshire.
80
3.2.2.3 Route 26 Bypass – New Gloucester, Maine
Several kilometers of road were constructed to bypass an unsafe section of Route
26 in New Gloucester, Maine. The subbase material consisted of 229 mm (9 in.) of
MaineDOT Type D aggregate overlying 229 mm (9 in.) of MaineDOT Type E aggregate.
Type D aggregate meets MaineDOT Specification 703.06 (152 mm (6 in.) maximum
size, 25 to 70% passing the 6.4 mm (¼ in.), 0 to 30% passing the No. 40, and 7%
maximum fines). Type E aggregate meets MaineDOT Specification 703.06 (152 mm (6
in.) maximum size, 25 to 100% passing the 6.4 mm (¼ in.), 0 to 50% passing the No. 40,
and 7% maximum fines).
Figure 3.14 Route 26 test section, New Gloucester, Maine.
The Type D aggregate was classified A-1-a according to the AASHTO
classification system. It has a maximum dry density of 1.99 Mg/m3 (124 lb/ft3) at an
81
optimum water content of 12%. The Type E aggregate was classified A-1-b according to
the AASHTO system. It has a maximum dry density of 1.94 Mg/m3 (121 lb/ft3) at an
optimum water content of 11%. Gradation and moisture density curves for both materials
are provided in Appendix A and B.
3.2.2.4 Route 201 Reconstruction – The Forks, Maine
A 20 km (12.41 mi) section of Route 201 in the town of The Forks, Maine, was
reconstructed during the summer of 2003 due to progressive pavement deterioration. The
existing pavement thickness ranged from 88.9 mm to 127 mm (3.5 in. to 5 in.). A cold in
place (CIP) technique was used to rehabilitate the existing pavement. The CIP process
begins by milling the existing pavement down. A 7 km (4.35 mi.) section was reclaimed
to a depth of 111 mm (3 in.), 8.2 km (5.08 mi) of 102 mm (4 in.), 3.8 km (2.35 mi) of 19
mm (0.75 in.) overlay, and 1 km (0.63 mi) of variable gravel. Each lane was completed
in one pass. The millings are conveyed to a crusher and passed over a 38 mm (1.5 in.)
screen. Material passing the screen is conveyed to a pugmill. Dry cement, water, and
emulsified asphalt are added in the following percentages based on dry weight: 0.5%
Portland cement; 2.8% water; and 1.7% emulsion. The material is mixed and conveyed
to a paver for laydown. Once the material was placed, it was rolled with a 10 ton steel
drum roller and a 20 ton pneumatic tired roller. The compaction process continued until
the pavement had a density of 98% of the target density. The target density of 2.0 Mg/m3
(126 lb/ft3) was established from a 91 m (300 ft) trial section. A 38 mm (1.5 in.) hot mix
asphalt (HMA) surface was placed over the CIP after a 10 day curing period. All testing
was done prior to the placement of the HMA surface.
82
3.2.2.5 Commercial Paving & Recycling – Scarborough, Maine
Commercial Paving & Recycling Co., Inc. (CPR) constructed a 67 m (220 ft) test
section at its recycling facility in Scarborough, Maine in October, 2003. CPR developed
a paving material known as Flexpave. This material is cold mix-asphalt made from waste
materials such as roofing shingles, bottom ash, fly ash, recycled glass, recycled asphalt
pavement (RAP), and virgin emulsified asphalt (MS-4). A 63 mm (2.5 in.) layer of
Flexpave was used to surface an existing gravel road and monitor its performance. The
mix used for the test section consisted of 61% roofing shingles, 30% bottom ash, and 9%
MS-4, by weight. Results from Marshall Stability tests on field samples are shown in
Table 3.14.
Table 3.14 Summary of Marshall Stability tests on field samples at Commercial Paving & Recycling test site, Scarborough, Maine.
Beginning of Test Section End of Test Section
Average Max. Load(kg)
Flow(%)
Average Max. Load(kg)
Flow (%)
252 15 210 18
The subgrade material was classified A-1-b according to the AASHTO system of
classification. The material has a maximum dry density of 2.05 Mg/m3 (128 lb/ft3) at an
optimum water content of approximately 6%. Gradation and moisture density curves are
provided in the Appendices.
83
3.3 INSTRUMENTATION Instrumentation was used for the spring thaw monitoring portion of this project to
quantify the condition of the test sections on days when measurements were made.
Instrumentation included: thermocouples, thermistors, frost tubes, vibrating wire
piezometers, standpipe piezometers, and TDR probes. Only selected instruments were
installed in each project as listed in Table 3.15.
Table 3.15 Summary of instruments spring thaw field test sites.
Field Test Site Instrumentation Kennebec Road Thermocouples, Standpipe Piezometers
Lakeside Landing Road Thermocouples, Standpipe Piezometers Stinson Lake Road Thermistors, Frost Tube, TDR probes
Buffalo Road Thermocouples, Frost Tube
USFS Parking Lot Thermistors, Standpipe Piezometers, TDR probes
Crosstown Road Thermocouples, Frost Tube Knapp Airport Parking Lot Frost Tube
Witter Farm Road Thermocouples, Standpipe Piezometer Route 126 Thermocouples, Vibrating Wire Piezometers
Route 1A Thermocouples, Standpipe Piezometers, Vibrating Wire Piezometers
3.3.1 Frost Penetration Measurement Thermocouples, thermistors, and/or frost tubes were installed and used to monitor
frost penetration during the winter and subsequent spring thaw. Thermocouples were
installed at Kennebec Road, Lakeside Landing Road, Buffalo Road, and Crosstown
Road. Thermistors were installed at the USFS Parking Lot site and at Stinson Lake
Road. Frost tubes were also installed at Buffalo Road, Crosstown Road, and Knapp
84
Airport Parking Lot. Thermocouple strings installed as part of previous projects were
used at the Witter Farm Road, Route 126, and Route 1A test sites.
3.3.1.1 Thermocouple Characteristics The thermocouples used for this project were 20 gage copper constantan (Type T). A bi-
metal reaction occurs where the copper and constantan wires are joined at their terminus
in the ground. Measurement of the resulting electrical potential allows the temperature to
be determined. The thermocouples installed for this project were attached to 2.4 m (8 ft)
long by 25 mm (1 in.) diameter wooden dowels. Eleven sensors were placed vertically
every 152 mm (6 in.) to a depth of 0.9 m (3 ft), where spacing was increased to 0.3 m (1
ft) for the remaining length of the probe. The twelfth sensor is a flyer, which is not
attached to the dowel and was placed in the soil directly beneath the pavement layer,
above the top of the dowel. Thermocouple string configuration is shown in Figure 3.15.
Thermocouple location at each site is provided in Table 3.16.
ASPHALT
SUBBASE
SUBGRADE
THERMOCOUPLE CABLE FLYER SENSOR
NOTES:1. NOT TO SCALE
Figure 3.15 Thermocouple string detail.
85
Table 3.16 Summary of thermocouple locations at spring thaw field test sites.
Test Site Test Section
Station (m)
Offset From C/L
(m)
Approximate Depth of Top Sensor
From Finish Grade mm (in.)
Approximate Depth of Flyer Sensor
From Finish Grade mm (in.)
1 NA 2.7 RT 305 (12) Kennebec Road 2 NA 2.7 RT 305 (12)
152 (6)
1 NA 2.7 RT 432 (17) 152 (6) Lakeside Landing
Road 2 NA 2.7 RT 203 (8) 152 (6)
Buffalo Road 1 NA 2.7 RT 238 (8) 152 (6)
Crosstown Road 1 NA 2.7 RT 152 (6) 102 (4)
Control 0+09.1 CL 248 (10) 1 0+70.1 CL 333 (13)
Witter Farm Road 2 0+57.9 CL 422 (17)
NFS
3 1+670 2.9 RT 305 (12) 8 4+040 2.9 RT 460 (18) Route 126
12 4+710 2.9 RT 152 (6) NFS
255+50 CL D-1 258+50 CL 267+50 CL D-2 268+50 CL
203 (8) Route 1A
E-3 88+85 7.5 RT 430 (17)
NFS
NFS – no flyer sensors.
3.3.1.2 Thermistor Characteristics
The thermistors used for this project are YSI epoxy encapsulated thermistors.
The temperature dependant resistors allow for a direct measurement of resistance. One
thermistor probe was installed in the USFS parking lot site in Rumney, N.H. in the fall of
2002. The probe is approximately 1.5 m (5 ft) long and consists of 8 thermistors. The
thermistors are located within the probe and are spaced roughly191 mm (7.5 in.) apart.
Three thermistors are flyers and are located above the top of the probe, at various depths
beneath the ground surface.
Two thermistor probes were installed at the Stinson Lake Road site during the
summer of 2003. Each probe is 1.6 m (5.25 ft) in length and contains 12 thermistors, 11
86
within the probe, and one flyer attached to the top of the probe. Two additional flyers
were installed with each thermistor string and are not attached to the probe. The
thermistors within the probes are spaced at 102 mm (4 in.) to a depth of 914 mm (35.5
in.) where spacing then increases to 152 mm (6 in.) for the remaining length of the probe.
Thermistor location at each site is shown in Table 3.17.
Table 3.17 Summary of thermistor locations at spring thaw field test sites.
Test Site Instrument No.
Offset From C/L
(m)
Approximate Depth of Top of Probe
From Finish Grade mm (in.)
Approximate Depth of Flyer Sensors
From Finish Grade mm (in.)
70 (3) 134 (5)
USFS Parking
Lot T1 NA 334 (13)
194 (8) 85 (3)
152 (6) T1 2.7 RT 393 (15) 186 (7) 79 (3)
165 (7)
Stinson Lake Road
T2 2.7 RT 405 (16) 195 (8)
NA – no centerline present in parking lot.
3.3.1.3 Frost Tube Characteristics
Frost tubes used for this project were approximately 1.8 m (6 ft) long, 18 mm (½
in.) plastic tubing. Within the plastic tubing is a blue colored fluid. When freezing
conditions exist, the blue color precipitates out leaving a colorless material. The depth of
frost penetration is determined by measuring the length of colorless material. The frost
tube is housed within a 19.1 mm (¾ in.) PVC pipe. The frost tubes used for this project
are shown in Figure 3.16. Frost tube locations are shown in Table 3.18.
87
Figure 3.16 Frost tube detail.
Table 3.18 Summary of frost tube locations at spring thaw field test sites.
Test Site InstrumentNo.
Offset From C/L (m)
Buffalo Road F1 2.7 RT
Knapp AirportParking Lot F1 NA
NA – no centerline in parking lot.
3.3.1.4 Installation & Monitoring
Thermocouple strings, thermistor probes, and/or frost tubes were installed in the
outside wheel path, approximately 2.7 m left or right of centerline as shown in Figure
3.17. A drill rig was required for installation. The drill rig augured a 102 mm (4 in.)
88
diameter hole to approximately 2.7 m (9 ft) below the road surface. Samples were taken
for water content, Atterberg Limit, and/or gradation laboratory tests as appropriate for the
soil type recovered. The strings/tubes were inserted and backfilled with native material
as shown in Figure 3.18.
2.7
2.4
ASPHALT
BASE COURSE
SUBBASE
SUBGRADE
1. ALL DIMENSIONS GIVEN ARE IN METERS.
COURSE
CL
THERMOCOUPLE CABLE
THERMOCOUPLE STRING
Figure 3.17 Thermocouple/thermistor section view.
A groove approximately 51 mm (2 in.) wide by 152 mm (6 in.) deep was jack-
hammered in the pavement to run the thermocouple/thermistor wire to the edge of the
pavement. After placement in the groove, the wire was covered with cold patch and
compacted. Beyond the edge of the pavement the wire was placed in PVC conduit and
buried in a 152 mm (6 in.) deep trench and run to the outside of the ditch line.
89
Figure 3.18 Placement of thermocouple string in auger hole.
Manual thermocouple readings were taken weekly during the spring thaw with a
hand held electronic readout unit (Omega Type HH201A). The accuracy of the readings
was improved by keeping the readout device at a constant temperature. This was done by
taking readings inside of a heated vehicle. The constant temperature is used a reference
temperature by the handheld unit to determine the temperature at the location of that
particular thermocouple pair. Lead wires from the thermocouples were installed with
adequate length to extend from the outside of the ditch line to the side of the road.
Thermocouple strings previously installed in Witter Farm Road, Route 1A, and
Route 126 are read automatically by a Campbell Scientific data acquisition system. The
thermistor probes installed to monitor the USFS sites are also read by automated data
acquisition systems. All data acquisition systems are read hourly for daily average
90
temperature determination. Data was downloaded approximately weekly during the
spring thaw. Frost tubes were also monitored weekly.
3.3.2 Pore Water Pressure Measurement
Standpipe piezometers were installed to monitor pore water pressures. At sites
where piezometers were not already present, four standpipe piezometers were installed.
In addition, time domain reflectometry (TDR) probes were installed in two USFS sites
under the direction of Maureen Kestler. Previously installed vibrating wire piezometers
were used in Route 1A and Route 126. Previously installed standpipe piezometers were
used at the Witter Farm Road and USFS Parking Lot.
3.3.2.1 Vibrating Wire Piezometer Characteristics
Vibrating wire piezometers were installed in Route 1A in the towns of Frankfort
and Winterport, Maine during the summer of 1997 and 1998. RocTest PWS vibrating
wire piezometers were used for this project. They have a low air entry sintered ceramic
porous stone and a 34 kPa (5 psi) range of measurement. Piezometers are located
beneath the break down lane, 3.7 m (12 ft) right of centerline. RocTest PWL vibrating
wire piezometers were installed in Route 126 in the towns of Monmouth and Litchfield,
Maine during the fall of 2001 and the summer of 2002. A summary of vibrating wire
piezometer locations is provided in Table 3.19.
91
Table 3.19 Summary of vibrating wire piezometer locations at spring thaw field test sites.
Test Site Test Section
Station (m)
Offset From C/L
(m)
Subbase or Subgrade
Approximate Depth From Finish Grade
mm (in.) 1 + 673 Subbase 405 (16) 3 1 + 673
3.7 RT Subgrade 660 (26)
4 + 042 Subbase 710 (28) 8 4 + 042 3.7 RT Subgrade 965 (38) 4 + 712 Subbase 255 (10)
Route 126
12 4 + 712 3.7 RT Subgrade 510 (20) Subbase 468 (18) Subgrade 950 (37) D-1 78 + 79 3.7 RT Subgrade NA Subbase 481 (19) Subgrade 715 (28) D-2 81 + 53 3.7 RT Subgrade 887 (35) Subbase NA
Route 1A
E-3 88 + 15 3.7 RT Subgrade NA NA – information not available.
3.3.2.2 Standpipe Piezometer Characteristics
For projects that did not already have piezometers, standpipe piezometers were
used. The piezometers are 25 mm (1 in.) diameter schedule 40 black iron pipe. The
piezometers extend approximately 76 mm (30 in.) into the road base material. The lower
300 mm (12 in.) of the pipe is slotted. Two pieces of 25 mm (1 in.) by 6.4 mm (¼ in.) by
300 mm (12 in.) steel were welded on opposing sides of the piezometer. This was done
to add resistance to aid in removing the top plug during measurement. The black iron
pipe was wrapped with geosynthetic fabric to prevent migration of soil particles into the
well. The top of the piezometer consists of a plug that is unscrewed to take
measurements. A detailed view of standpipe piezometers is shown in Figure 3.19.
92
Standpipe piezometers used at the Witter Farm Road are shown in Figure 3.20.
Piezometer location is shown in Table 3.20.
254 mm
102 mm
659 mm
25 mm
BACKFILLED
BENTONITE SEAL
CONCRETE SAND
CONCRETE SAND
102 mm
25.4 mm BLACK
WELDED FINS
6.35 mm DIA. HOLES
25.4 mm BLACK IRON
25.4 mm BLACK
25.4 mm BLACKNATURAL SOILIRON PLUG
IRON PIPE
COUPLING
IRON CAP
Figure 3.19 Standpipe piezometer detail.
Figure 3.20 Standpipe piezometer detail for Witter Farm Road, Orono, Maine (Lawrence, et al., 2000).
93
Table 3.20 Summary of standpipe piezometer locations at spring thaw field test sites.
Test Site Test Section
Station (m)
InstrumentNo.
Offset FromC/L
(m)
Approximate Depth of Top of Piezometers
From Finish Grade mm (in.)
P1 1 NA P2 P3
Kennebec Road
2 NA P4
2.7 RT 76 (3)
P1 1 NA P2 P3
Lakeside Landing
Road 2 NA P4
2.7 RT 152 (6)
1 0+70.1 P1 LT Control 0+3.1 P2 RT
Witter Farm Road
5 0+18.3 P3 LT NA
255+50 P1 13.1 LT 256+75 P2 11.6 RTRoute 1A D-1 261+00 P3 10.8 LT
NA
NA – information not available.
Standpipe piezometers were installed at the USFS parking lot site in Rumney,
N.H. in the fall of 2002. They extend to approximately 1.8 m (6 ft) below the ground
surface. This type of piezometer consists of a filtered tip connected to a riser pipe. The
standpipe piezometers are 3 m (10 ft) long, 51 mm (2 in.) diameter PVC pipe. The
perforated section is covered with geosynthetic fabric to prevent migration of particles
into the well. The riser pipe is terminated above the ground surface and covered with a
vented cap.
3.3.2.3 TDR Probe Characteristics
The TDR probes used for this project are Soilmoisture 6005L2 Buriable
Waveguides. A frequency electromagnetic pulse is generated and sent down a line
94
comprised of two waveguides. The velocity of propagation of the high frequency, broad
band 3GHz wave in soil is determined primarily by the water content. The wave is
reflected from the open ends of the waveguides and returns along the original path. By
microprocessor, the travel time of the wave is used to directly calculate the dielectric
constant of the soil. TDR probe locations are shown in Table 3.21.
Table 3.21 TDR probe locations at Stinson Lake Road and USFS Parking Lot, Rumney, New Hampshire.
Test Site InstrumentNo.
Offset FromC/L
(m)
Approximate Depth of Probe From Finish Grade
mm (in.) TDR 1 677 (27) TDR 2 561 (22) TDR 4 427 (17) TDR 5 268 (11)
Stinson Lake Road
TDR 6
2.7 RT
143 (6) TDR 1 NA USFS Parking
Lot TDR 2 NA NA – information not available.
3.3.2.4 Installation & Monitoring
Installation of vibrating wire piezometers used at Routes 1A and 126 and
standpipe piezometers used at the Witter Farm Road and USFS Parking Lot are not
discussed.
The standpipe piezometers were installed in the outside wheel path,
approximately 2.7 m (9 ft) left or right of centerline. A drill rig was required for
installation. The drill rig augured a 102 mm (4 in.) diameter hole. The piezometers were
inserted into the augured hole and backfilled according to the specification illustrated in
95
Figure 3.19. Bentonite powder/pellets were used to minimize infiltration of surface
water.
All monitoring was done approximately weekly during the spring thaw. The top
plug was removed and the depth to the water surface from the top of pavement was
measured and recorded. Vibrating wire piezometers used in Section 3 and Section 8 on
Route 126 are read automatically by a Campbell Scientific data acquisition system.
Vibrating wire piezometers in Section 12, Route 126, and those used at Route 1A were
read manually with a RocTest MB-6T unit. TDR probes were read with a Trase system
6050X1 measuring unit.
3.4 FIELD TESTING PROCEDURES
3.4.1 Spring Thaw Monitoring
At each field site, device measurements were performed at a minimum of eight
locations. Measurements were taken approximately weekly during the spring thaw
period. Test point locations for Kennebec Road and Lakeside Landing Road field sites
are shown in Figures 3.21. Test point locations for Stinson Lake Road, Buffalo Road,
and Crosstown Road are shown in Figure 3.22. Test point locations for USFS Parking
Lot and Knapp Airport Parking Lot are shown in Figure 3.23. Table 3.22 provides test
point locations for Witter Farm Road, Route 126, and Route 1A.
96
2.70
14.642.44 2.44
14.642.442.44
NOTES:1. ALL DIMENSIONS SHOWN ARE IN METERS
LEGEND:= FWD & PFWD TEST POINT LOCATION= STANDPIPE PIEZOMETER
= THERMOCOUPLE STRING
SECTION 1SECTION 2
TP1TP2TP3TP4 TP1TP2TP3TP4
EXISTING CENTERLINE OF ROAD
Figure 3.21 Kennebec Road and Lakeside Landing Road test point layout.
2.70
EXISTING CENTERLINE OF ROAD
NOTES:1. ALL DIMENSIONS SHOWN ARE IN METERS
LEGEND:= FWD & PFWD TEST POINT LOCATION
TP1TP2TP3TP4TP5TP6TP7TP8TP9TP10
Figure 3.22 Stinson Lake Road, Buffalo Road, and Crosstown Road test point layout.
97
Table 3.22 Summary of test point locations at Witter Farm Road, Route 126, and Route 1A.
Witter Farm Road Route 126 Route 1A
Test Section
Station (m)
Offset From C/L (m)
Test Section
Station (m)
Offset From C/L (m)
Test Section
Station (m)
Offset From C/L (m)
1.2RT 1+652 255+50 0+3 0.6LT 1+664 256+50 0.6RT 1+676 257+50 0+6 1.2LT
Section 3
1+688 258+50 1.2RT 4+028 259+00
Control Section
0+9 0.6LT 4+036
Section D-1
260+50 0.6RT 4+044 262+00 0+55 1.2LT
Section 8
4+052 263+00 1.2RT 4+704 264+00 0+58 0.6LT 4+708 265+00 0.6RT 4+712 266+00
Section 2
0+61 1.2LT
Section 12
4+716
2.7RT
Section D-2
267+00 0.6RT 291+00 0+67 1.2LT 292+00 1.2RT 293+00 0+70 0.6LT 294+00 0.6RT
Section E-3
295+00
2.7RT
Section 1
0+73 1.2LT
3.4.1.1 Portable Device Measurements
Several portable measuring devices were used for testing during the spring of
2004. Prima 100 PFWD measurements were taken at all test sites. Loadman PFWD
measurements were taken at spring thaw test sites in Rumney, New Hampshire. Clegg
Impact Hammer and Humboldt Soil Stiffness Gauge measurements were taken at the
USFS Parking Lot during the spring of 2003 and 2004. Each device is discussed
separately in the following sections.
3.4.1.1.1 Prima 100 PFWD
The Prima 100 is a light, portable device that has been developed to measure
stiffness of construction layers including subgrades, base courses, and pavements.
98
Details regarding the mode of operation are provided in Section 2.2.2.1. With the Prima
100 PFWD, six measurements were taken at each of three different drop heights, at each
test location. Additional drop heights were approximately equal to 850, 630, and 420
mm (34, 25, and 17 in.). This is illustrated in Figure 3.23. Additional deflection sensors
were used with spacing as follows (as measured from the center of the loading plate): 0,
207, and 407 mm (0, 8, and 16 in.). The PFWD measurements were taken utilizing a 20
kg (44 lb) drop weight and a 300 mm (11.8 in.) loading plate. In all cases, the first
reading was neglected and the average of the remaining five was used for analysis and
comparison. These, as well as other input parameters are summarized in Table 3.23.
(a) (b) (c)
Figure 3.23 Variable drop heights (a) 850 mm, (b) 630 mm, and (c) 420 mm.
Table 3.23 Prima 100 PFWD input parameters.
99
Setup Menu Input AsphItem Parameter
alt SurfacedTest Sites
Gravel SurfacedTest Sites
Pretr s) *ig time (m 10 Pulsebase (%) 24* Trigger
0.90* Trig Level (kN) View Sample Time (ms) 60*
Load Plate Radius (mm) 150 Number of sensors 3
D(1) offset (cm) 0 D(2) offset (cm) 20.7
Mechanical
40.7 D(3) offset (cm) Poisson’s Ratio 0.35** 0.35** Formula
Stress Distribution 2.0 2.67 * - default values. ** - Huang, 2004.
.4.1.1.2 Loadman PFWD
of the Loadman PFWD and its mode of operation is
provide
drops
.4.1.1.3 Clegg Impact Hammer
e Clegg Impact Hammer and its mode of operation is
provide
e
Eqn. 3.1
Where: E Elas Mod (MPa) CIV = Clegg Impact Value
3
A detailed description
d in Section 2.2.2.2. A total of five measurements were taken at each test
location. In all cases, the first drop was neglected and an average of the remaining
was used for analysis and comparison.
3
A detailed description of th
d in Section 2.2.1. Four drops were made at each test location. The first drop
was excluded and the average of the remaining drops was averaged. The average of th
values was taken and correlated to stiffness using Equation 3.1. This value was used for
analysis and comparison.
2088.0 CIVE ⋅=
= tic ulus
100
3.4.1.1.4 Humboldt Soil Stiffness Gauge
The Humboldt Soil Stiffness Gauge (SSG) was used for testing at the USFS
Parking Lot site in Rumney, New Hampshire during the spring of 2003. A detailed
description of the SSG is provided in Section 2.2.1. One measurement was taken at each
test location.
3.4.1.2 Falling Weight Deflectometer Testing
The MaineDOT provided a falling weight Deflectometer (FWD) for seasonally
posted roads in Maine. The United States Army Corps of Engineers Cold Regions
Research and Engineering Laboratory (CRREL) provided a FWD for test sites in
Rumney, New Hampshire. The Vermont Agency of Transportation (VAOT) provided a
FWD for seasonally posted low volume roads in Vermont. Each agency utilized different
units and had different testing techniques. As a result, each is discussed separately
below.
3.4.1.2.1 MaineDOT FWD
The MaineDOT utilizes a JILS Model 20C Falling Weight Deflectometer
manufactured by Foundation Mechanics, Inc. The unit has a constant drop weight of
340.2 kg (750 lb). The load capacity ranges from 9 to 120 kN (2,000 to 27,000 lbf) with
a loading plate diameter of 304.8 mm (12 in.). Deflection sensors are spaced at 0, 305,
457, 610, 914, 1219, and 1524 mm (0, 12, 18, 24, 36, 48, and 60 in.). The unit is shown
in Figure 3.24.
101
Figure 3.24 MaineDOT JILS Model 20 C FWD.
force calibration. During the
alibration, the software determines the drop heights required to produce predetermined
The
CRREL uses a Dynatest Model 8000 Falling Weight Deflectometer. The Model
range of 7 to 120 kN (1.5 to 27.0 kips). The unit has a loading
late di
ts.
Prior to testing the FWD operator conducts a
c
forces based on layer response. One drop each at six different loads is performed.
loading sequence is as follows: 26.7, 40.0, 53.4, and 71.2 kN (6, 9, 12, 16, 9, and 9 kips).
3.4.1.2.2 CRREL FWD
8000 FWD has a loading
p ameter of 457 mm (18 in.). Deflection sensors are spaced at 0, 305, 610, 914,
1219, 1524, and 1829 mm (0, 12, 24, 36, 48, 60, and 72 in.). The unit is shown in Figure
3.25. The CRREL test program targeted four drops at each of four different drop heigh
102
Figure 3.25 CRREL Dynatest 8000 FWD.
The VAOT utilizes a Dynatest® Model 8000 Falling Weight Deflectometer. The
loading range of 7 to 120 kN (1.5 to 27.0 kips). The unit has a
3.4.1.2.3 VAOT FWD
Model 8000 FWD has a
loading plate diameter of 300 mm (11.8 in.) and deflection sensor spacing of 0, 203, 305,
457, 610, 914, 1219, 1524, and 1829 mm (0, 8, 12, 18, 24, 36, 48, 60, 72 in.) The unit is
shown in Figure 3.26.
103
Figure 3.26 VAOT Dynatest 8000 FWD.
VAOT FWD testing followed the Strategic Highway Research Program (SHRP)
FWD testing protocol administered by the Long Term Pavement Performance (LTPP)
division of the Federal Highway Administration (FHWA). This testing procedure
includes three seating drops and four drops each at four different drop heights that target
four different loads. The procedure is shown in Table 3.24 and Table 3.25.
Table 3.24 FLEX testing plan drop sequence used at Berlin, Vermont test sites (LTTP, 2000).
No. of Drops Height (mm) Data Stored
3 3 (200) No*
4 1 (50) Peaks 4 2 (100) Peaks 4 3 (200) Peaks 4 4 (390) Peaks & History
* - no data stored, seating drop only. Deflection and load data are printed but not stored.
104
Table 3.25 FLEX testing plan target loads used at Berlin, Vermont test sites (LTTP, 2000).
Height (mm)
Target LoadkN (kips)
Acceptable Range kN (kips)
1 (50) 27 (6.0) 24.0 to 29.4 (5.4 to 6.6) 2 (100) 40 (9.0) 36.0 to 44.0 (8.1 to 9.9) 3 (200) 53 (12.0) 48.1 to 58.7 (10.8 to 13.2)4 (390) 71 (16.0) 64.1 to 78.3 (14.4 to 17.6)
3.4.2 Subgrades and Construction Materials
The field component included tests on two subgrades, one construction sand
product, two aggregates, and one reclaimed stabilized base product. At each field site,
tests were performed utilizing both the Prima 100 PFWD and Nuclear Moisture Density
Gauge (NDM) (AASHTO T 238). Multiple tests were performed at a minimum of 12
locations using each instrument. Test point locations for measurements in Southington,
Connecticut were similar to those depicted in Figure 3.32. Test point locations for all
other sites are shown in Figure 3.27. Samples were taken at each site for sieve analysis,
maximum dry density, and optimum water content determination.
105
= PFWD & NDM TEST POINT LOCATION
1. ALL DIMENSIONS SHOWN ARE IN METERS
LEGEND:
NOTES:
TP3
APPROXIMATE CENTERLINE OF ROAD
TP5 TP4 2.70TP2 TP1TP1TP2TP3TP4TP5TP1TP2TP3TP4TP5
2. NOT TO SCALE
Figure 3.27 Test point layout for compaction control field test sites.
With the Prima 100 PFWD, six measurements were taken at each test location.
The maximum drop height of 850 mm (33.5 in.) was used throughout. Setup input
parameters are shown in Table 3.23. In all cases, the first reading was neglected and the
average of the remaining five was used for analysis and comparison.
For field test sites in Maine, NDM measurements were taken with a MC-1
Portaprobe manufactured by Campbell Pacific Nuclear International. The device is
shown in Figure 3.28. NHDOT and Connecticut Department of Transportation provided
NDM’s for field sites in their respective states. Both departments provided Troxler 3430
gauges. This device is shown in Figure 3.29. Five NDM measurements were taken at
depths of 203, 152, 102, 51, and 0 mm (8, 6, 4, 2, and 0 in.) at each test location. Density
measurements taken at 203 mm (8 in.) depths were used for analysis and comparison.
106
Water content values determined from surface measurements were used for analysis and
comparison.
Figure 3.28 CPN MC-1 Portaprobe NDM.
Figure 3.29 Troxler 3430 NDM.
107
3.5 LABORATORY TESTING PROCEDURES
3.5.1 Spring Thaw Monitoring
During instrumentation installation samples were taken for water content,
Atterberg Limit, and/or gradation laboratory tests as appropriate for the soil type
recovered. Laboratory tests followed procedures outlined by AASHTO.
3.5.2 Subgrades and Construction Materials
The primary purpose of the laboratory component of this project is to determine a
relationship between PFWD results and percent compaction under carefully controlled
conditions. Five different material types were used. Each aggregate was compacted in
the container to approximately 90, 95, and 100% of the maximum dry density. The effect
of water content was determined at 95% of the maximum dry density. Measurements
were taken at optimum water content as well as ± 3% of the optimum water content.
Layer construction and testing procedures are discussed below.
3.5.2.1 Test Section Construction
The large-scale laboratory study to correlate PFWD results to percent compaction
was constructed in the geotechnical research laboratory at the University of Maine. The
tests were conducted in a 1.8 m x 1.8 m x 0.9 m (6 ft x 6 ft x 3 ft) deep test container as
shown in Figure 3.30.
108
Figure 3.30 Laboratory test box.
The bottom 203 mm (8 in.) of material met MaineDOT Type D Aggregate
specifications and was left in-place throughout the testing. This base layer was placed in
two, approximately equal lifts and compacted to 100% of the maximum dry density and
optimum moisture content. To differentiate between the base layer and the overlying
layers which would be changed with each test, geogrid and/or geotextile were added on
top of the lower layer. Material was then added in three 152 mm (6 in.) lifts. NDM
readings were taken at 152 mm (6 in.) depths at two to three locations after each lift was
added in order to determine whether predetermined compaction and water content
requirements had been met. Each lift was compacted in the container to approximately
90, 95, and 100% of the maximum dry density (AASHTO T 180). Compaction was
achieved by means of a Bosch 11304 Brute Breaker Hammer. Two different tamper
109
plates were used; both were Bosch HS2124 152 mm (6 in.) square tamper plates. One of
the plates was modified; a 305 mm x 305 mm x 6 mm (12 in. x 12 in. x ¼ in.) steel plate
was welded to the smaller tamper plate. The outfitted jackhammer with modified flat
plate attachment is shown in Figure 3.31.
Figure 3.31 Bosch 11304 hammer with modified flat plate attachment.
3.5.2.2 Portable Device Measurements
Once construction of the test sections was completed multiple portable devices
were used. Devices used for testing included Prima 100 PFWD, Clegg Impact Hammer,
NDM, and Dynamic Cone Penetrometer (DCP). Measurements were taken at five
locations in a pattern similar to the one shown in Figure 3.32.
110
= WATER CONTENT LOCATION
1. ALL DIMENSIONS GIVEN IN METERS
LEGEND
NOTES:1.83
1.83
0.46
0.46
0.46
0.46
0.46
= NDM, CLEGG, AND Prima 100 PFWD TEST POINT LOCATION
= DCP TEST POINT LOCATION
= SAND CONE TEST LOCATION
Figure 3.32 Laboratory test point layout.
Prima 100 and Clegg Impact Hammer measurements were taken following the
procedures outlined in Sections 3.4.1.1.1 and 3.4.1.1.3. NDM measurements were taken
following procedures outlined in Section 3.4.2.1 with a MC-1 Portaprobe NDM. Two
water content samples were taken from the surface at locations indicated in Figure 3.27
and used to compare with those obtained with the NDM. One sand cone test was
completed for each trial to compare with densities obtained with the NDM. Lastly, two
dynamic cone penetrometer (DCP) tests were completed for each trial and followed
procedures outlined by AASHTO.
111
3.6 SUMMARY
The performance of seven paved and three gravel surfaced roads were monitored
during the spring of 2004 to evaluate the effectiveness of the Prima 100 PFWD in
tracking seasonal stiffness variations. Field test sites were located in Maine, New
Hampshire, and Vermont. Some sites were part of previous or ongoing NETC research
projects. The remainder of the sites were chosen by the researchers with aid from
MaineDOT and VAOT. Test sites were instrumented with thermocouples, thermistors,
and frost tubes to monitor subsurface temperatures. Vibrating wire and standpipe
piezometers as well as TDR probes were installed to monitor pore water pressure in the
subbase and subgrade layers and water content. Instruments were read manually
approximately weekly during the spring thaw period. Selected sites contained automated
data acquisition systems which monitored hourly. Data was downloaded approximately
weekly. Traditional and portable FWD’s, as well as other portable devices were used at
multiple locations at each test site through the spring and into early summer of 2004.
Five field sites were used for the evaluation of subgrades and construction
materials. Different aggregate types were tested at each field site. Field test sites were
located in Maine, New Hampshire, and Connecticut. Sites were located with the
assistance of the NETC Technical Committee assigned to this project and MaineDOT.
Prima 100 PFWD and NDM measurements were taken at multiple locations at each site.
Laboratory tests were performed on five different material types representative of
typical New England subbase materials. Each lift was compacted in the container to
approximately 90, 95, and 100% of the maximum dry density. The effect of water
content was determined at 95% of the maximum dry density. Measurements were taken
112
at optimum water content as well as ± 3% of the optimum water content. Each material
was added to the test box in approximately equal lifts. Materials were compacted with a
hand tamper and electric jackhammer until the predetermined compaction criteria had
been met. Prima 100 PFWD, Clegg Impact Hammer, NDM, and DCP measurements
were taken at multiple locations for each trial and material tested.
113
CHAPTER 4
SPRING THAW MONITORING
4.1 INTRODUCTION
This chapter presents the analysis and results of monitoring seasonal stiffness
variations in paved and unpaved, seasonally posted, low volume roads. The objective of
this portion of the research project was to investigate the ability of the Prima 100 Portable
Falling Weight Deflectometer (PFWD) to track seasonal stiffness variations.
Comparisons were made to the traditional Falling Weight Deflectometer (FWD) as well
as other portable devices. Correlations were developed to compare performance.
Recommendations are made for field testing techniques.
The performance of seven paved and three gravel surfaced roads were evaluated
during the spring and early summer of 2004. Portable and traditional FWD tests were
performed at multiple locations at each site beginning in early March with the last set of
readings taking place in late June. Additional measurements were taken at the United
States Forest Service (USFS) Parking Lot during the spring of 2003. One set of
measurements was taken on Route 11, Wallagrass Plantation, Maine and Route 167,
Presque Isle/Fort Fairfield, Maine during the spring of 2003. These results are presented
separately.
This chapter is organized as follows. Frost penetration and pore water pressure
measurements are presented first, followed by seasonal stiffness variations measured with
the FWD and PFWD. Portable and traditional FWD derived moduli are compared.
Comparisons are made to other portable measuring devices used at selected sites.
114
Finally, field testing techniques are evaluated and appropriate recommendations are
presented.
4.2 FROST PENETRATION
Subsurface temperatures were measured at each field site during the end of the
freezing season, throughout the thawing period, and into the recovery period. These
results are presented first to provide the context for interpretation of PFWD and FWD
results in subsequent sections. Temperature readings were generally taken weekly,
however, at a few sites readings were taken hourly by an automatic datalogger. Details
of the instrumentation installed to measure subsurface temperatures are given in Section
3.3.1.
The subsurface temperature measurements were used to determine the maximum
depth of the 0˚C (32°F) isotherm. This was assumed to be the maximum depth of frost
penetration, ignoring factors such as the salinity of the porewater that can alter the
freezing point of water. Thawing occurs both from the surface down and bottom up,
although the former tends to be the dominant factor. The initiation of surface thawing
was indicated by the first date that the temperature sensor closest to the surface had a
reading above 0˚C (32°F). The surface may undergo several cycles of thawing and
refreezing during the thawing season in response to daily and diurnal temperature
fluctuations. When all temperature sensors in the vertical string were above 0˚C (32°F),
it was assumed that the location was completely thawed. Since the temperature readings
at most of the sites were taken weekly, the dates of maximum depth of frost penetration,
115
initiation of surface thawing, and completion of thawing could only be determined
approximately.
The maximum depth of frost penetration, initiation of surface thawing, and the
completion of thawing are summarized in Table 4.1 for paved sites, and Table 4.2 for
gravel surfaced sites. Measurements taken at asphalt surfaced test sites indicated freezing
temperatures penetrated to their maximum depths between February 17 and March 24.
Maximum depths ranged from a minimum of 866 mm (34 in.) at Stinson Lake Road to a
maximum of 1930 mm (76 in.) at Route 1A (Section D-2). Complete thaw occurred at all
test sites between mid-March and mid-April.
Table 4.1 Summary of frost penetration measurements made on asphalt surfaced test sites.
Field Test Site
Test Section
Date of Maximum
Frost Penetration
Depth of Maximum
Frost Penetration
mm (in.)
First Day Top
Sensor Reads >0°C
Date of Complete
Thaw
1 3/24/2004 1852 (73) 3/12/2004 4/17/2004 Kennebec Road 2 3/10/2004 1372 (54) 3/12/2004 4/17/2004
Buffalo Road 1 3/11/2004 1846 (73) 3/3/2004 4/8/2004
Stinson Lake Road 1 2/28/2004 866 (34) 3/1/2004 4/19/2004
Knapp Airport Parking Lot 1 3/4/2004 1372 (54) NA 4/16/2004
3 3/17/2004 1214 (48) 3/26/2004 4/14/2004 Route 126 8 2/17/2004 1158 (46) 3/21/2004 4/7/2004
Control 3/1/2004 1594 (63) 3/3/2004 4/20/2004 2 2/20/2004 1135 (45) 3/15/2004 3/25/2004
Witter Farm Road
1 2/26/2004 1518 (60) 3/16/2004 4/16/2004 D-1 3/12/2004 1575 (62) 3/16/2004 3/30/2004 D-2 3/12/2004 1930 (76) 3/12/2004 3/16/2004 Route 1A E-3 3/2/2004 1725 (68) 3/16/2004 3/23/2004
NA – not available, frost tube measurements made.
116
Measurements taken at gravel surfaced test sites indicated freezing temperatures
penetrated to their maximum depths between March 1 and April 21. Maximum depths
ranged from a minimum of 1128 mm (44 in.) at the USFS Parking Lot to a maximum of
2134 mm (84 in.) at Crosstown Road. Complete thaw had occurred at all sites between
early April and mid May.
Table 4.2 Summary of frost penetration measurements made on gravel surfaced test sites.
Field Test Site
Test Section
Date of Maximum
Frost Penetration
Depth of Maximum
Frost Penetration
mm (in.)
First Day Top
Sensor Reads >0°C
Date of Complete
Thaw
1 4/21/2004 1803 (71) 3/12/2004 5/11/2004 Lakeside Landing
Road 2 3/16/2004 1422 (56) 3/2/2004 4/6/2004
USFS Parking Lot 1 3/1/2004 1128 (44) 2/22/2004 4/23/2004
Crosstown Road 1 4/16/2004 2134 (84) 4/2/2004 5/14/2004
4.3 PORE WATER PRESSURE
Pore water pressures were measured in the subbase layer at each field site
throughout the thawing and into the recovery period. At some sites, pore water pressures
were also measured in the subgrade layer. These results, like subsurface temperatures,
are provided for interpretation of the PFWD and FWD results presented in the following
sections. Pore water pressure measurements were taken approximately weekly, however,
at some sites readings were taken hourly by an automated data acquisition system.
117
Details of the instrumentation installed to measure subbase and subgrade pore water
pressures are provided in Section 3.3.2.
The pore water pressure measurements were used in conjunction with the
subsurface temperature readings to examine the extent to which the road had thawed and
recovered. Water, located in the pore space between soil particles, freezes as heat is
removed from the soil. The ice crystals grow by incorporating nearby water into the
crystal as more heat is removed. Capillary action draws water from the groundwater
table to the freezing front (0°C (32°F) isotherm). The ice crystals grow and merge
together to form ice lenses. This process reduces the density of the soil as it expands to
make room for the ice lenses, also creating frost heaves. This is illustrated in Figure 4.1.
Thawing predominantly occurs from the ground surface downward. Once some surface
thawing has occurred, water may be trapped above the underlying soil that is still frozen
and is not able to drain. As a result, a temporary loss of bearing capacity occurs. An
undrained loading condition can be created from passing traffic (Janoo and Cortez, 2002).
This phenomenon is known as thaw weakening. Once the entire road section has thawed,
water is able to drain and strength is regained, this is known as the recovery period.
118
Figure 4.1 Formation of ice lenses within a pavement structure (WSDOT). Manual standpipe piezometer readings are provided in Table 4.3. Two of four
standpipe piezometers installed at Kennebec Road became inoperable in late spring. The
reasons for this are that the top plug jammed on one and could not be removed, while
frost action heaved the top of the other up above the road surface and it was clipped off
by a snow plow. A summary of time domain reflectometry probe readings is provided in
Table 4.4. Manual vibrating wire piezometer measurements are presented in Table 4.5.
Figures 4.2 and 4.3 present the pore water pressure readings taken by the automated data
acquisition system at Route 126, Monmouth/Litchfield, Maine.
119
The highest water levels observed in standpipe piezometers roughly correspond to
the date of complete thaw for that particular site. This is true for both asphalt and gravel
surfaced test sections. Time Domain Reflectometry (TDR) moisture sensors used at the
USFS Parking Lot and Stinson Lake Road also indicate an increase in moisture content at
or near the date of complete thaw.
Table 4.3 Summary of standpipe piezometer measurements.
Kennebec Road Lakeside Landing Road
Witter Farm Road Route 1A Date
P1 P2 P3 P4 P1 P2 P3 P4 P1 P2 P3 P1 P2 P3 3/2 0.9N 0.9N 0.4F 0.4F 0.9N 0.9N 1.0N 1.0N 1.9 2.2 2.1 0.2 3.2 1.3 3/10 0.9F 0.9N 0.4F 0.4F 0.9N 0.9N 1.0N 1.0N 0.5 0.7 1.5 NA NA NA3/12 0.9F 0.9N 0.4F 0.4F 0.9N 0.9N 1.0N 0.9N 0.4 0.6 1.3 0.2 2.7 1.0 3/16 NA NA NA NA 0.9N 0.9N 0.6 0.8 0.5 0.6 0.9 0.2 2.6 0.8 3/19 0.9F 0.9N 0.4F 0.4F NA NA NA NA 0.5 0.6 1.1 0.2 2.5 1.0 3/23 NA NA NA NA NA 0.6 0.6 0.4 0.5 1.0 0.2 2.5 0.9 3/24 NA NA 0.4 NA NA NA NA NA NA NA NA NA NA3/30 NA NA NA 0.4 0.4 0.4 0.3 0.4 0.5 0.6 0.2 2.1 0.8 3/31 0.4 0.4 0.2 NA NA NA NA NA NA NA NA NA NA4/6 NA NA NA 0.4 0.3 0.3 0.4 0.4 0.5 0.5 0.3 2.0 0.8 4/7 0.4 0.4 0.1 NA NA NA NA NA NA NA NA NA NA4/17 0.9 0.2 NA NA NA NA NA NA NA 1.0 1.8 0.9 4/18 NA NA 0.1 0.1 0.3 0.4 NA NA NA NA NA NA4/21 NA NA 0.1 0.2 0.3 0.4 0.5 0.5 0.6 0.7 1.7 1.0 4/22 N 0.2 NA NA NA NA NA NA NA NA NA NA4/27 NA NA 0.1 0.1 NA 0.5 0.5 0.5 0.6 0.7 1.8 1.2 4/28 N 0.4 NA NA NA NA NA NA NA NA NA NA5/11
Inop
erab
le
NA
Inop
erab
le
NA 0.9 0.9 NA NA 0.7 0.6 0.5 NA NA NA*** Values indicate depth (m) to water in piezometer measured from the ground surface. N = no water, F = frozen, NA = not available
Table 4.4 Summary of time domain reflectometry probe water content readings.
USFS Parking Lot Stinson Lake Road Date TDR 1 (%) TDR 2 (%) TDR 1 (%) TDR 4 (%) TDR 5 (%) TDR 6 (%)
3/3 4.0 7.8 9.5 12.7 20.7 29.8 3/11 6.0 8.8 11.7 23.7 19.6 24.2 3/18 12.9 8.5 13.5 25.1 18.6 20.2 3/25 12.8 8.2 14.8 25.8 16.8 19.4 4/1 24.6 28.6 17.1 NA NA NA 4/8 19.6 15.8 27.7 33.5 30.3 27.4 4/15 20.8 13.6 34.6 33.7 30.3 26.9 4/29 13.3 12.4 34.8 33.0 29.5 20.9 5/13 10.5 11.6 NA NA NA NA 6/9 11.2 10.2 32.3 24.4 30.9 16.6
NA – not available.
120
The relationship between manual vibrating wire piezometer measurements and the date
of complete thaw is less clear. Manual readings do not show a trend. Manual readings
represent the conditions at the time the measurements were taken whereas automated
readings were taken hourly and averaged over a 24 hour period. Trends exhibited by
automated vibrating wire piezometer measurements roughly correspond with partially
and completely thawed states. Pore water pressure measurements during the frozen state
provide no meaningful information. Converting maximum subbase and subgrade pore
water pressure measurements into feet of head indicates that these readings are
unreasonable as the water surface lies above the finished road surface. Prior to the onset
of thaw, pressure reduces to roughly 2 kPa (0.67 ft of water) in Section 3 and 1 kPa (0.33
ft of water) in Section 8. Subbase pore water pressure remains relatively constant
through the end of the monitoring period. Subgrade pore water pressure in both sections
begins to increase just prior to the beginning of thaw and continue to increase to a
maximum value of nearly 20 kPa (6.6 ft of water) at the end of the monitoring period.
Table 4.5 Summary of manual vibrating wire piezometer measurements.
Pore Water Pressure (kPa)* Route 1A Route 126
Section D-1 (STA 258+50) Section D-2 (STA 267+25) Section 12 (STA 4+712) Date subbase subgrade subbase subbase
Date subbase subgrade
3/2 5.88 6.17 -1.86 1.78 3/27 -0.02 0.41 3/12 4.65 5.73 3.99 1.80 3/31 -0.15 0.46 3/16 4.77 5.54 2.45 1.20 4/16 -0.07 0.51 3/19 4.97 7.22 1.46 0.67 4/28 -0.02 1.19 3/23 4.71 7.97 1.65 0.87 5/12 0.14 1.04 3/30 4.28 6.05 2.28 0.71 4/6 4.76 3.92 2.72 0.85 4/17 4.66 3.79 1.81 1.46 4/21 4.55 1.65 0.00 1.52 4/27 4.39 4.34 1.63 1.27 * - 1 kPa ≈ 0.33 ft of water.
121
1/1/041/19/04
2/6/042/24/04
3/13/043/31/04
4/18/045/6/04
5/24/046/11/04
6/29/04
Date
-6
-4
-2
0
2
4
6
Hea
d of
Wat
er (f
t)
Frozen Partially Frozen
Thawed
Subgrade
Subbase
Figure 4.2 Route 126 (Section 3), Monmouth/Litchfield, Maine automated pore water pressure measurements.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
-6
-4
-2
0
2
4
6
Hea
d of
Wat
er (f
t)
Subbase
Subgrade
Frozen ThawedPartially Frozen
Figure 4.3 Route 126 (Section 8), Monmouth/Litchfield, Maine automated pore water pressure measurements.
122
4.4 SEASONAL STIFFNESS VARIATIONS
The seasonal variation in modulus as measured by the conventional and portable
FWDs are presented and assessed in this section. Details regarding backcalculation
analysis of FWD data are presented first, followed by the presentation and discussion of
results for asphalt surfaced roads. Lastly, the results derived from testing on gravel
surfaced roads are presented.
4.4.1 Backcalculation of Layer Moduli
Backcalculation is the process by which pavement layer moduli are determined by
matching measured and calculated surface deflection basins (FHWA, 1999).
Backcalculation of pavement layer moduli was performed for FWD test results and used
as the basis for comparison for the PFWD. MaineDOT provided FWD backcalculation
analysis for Maine test sites using DARWin. In addition, backcalculation was performed
on FWD data from all test sites using Evercalc. Details are discussed below.
4.4.1.1 Mid-Depth Asphalt Temperature Determination
Deflection measurements taken on all pavements are dependent on seasonal
variations that affect the underlying aggregate and subgrade. The results from asphalt
pavements are also dependent on the temperature of the asphalt. In order to meaningfully
analyze the deflection results, the deflections or deflection analysis results, must be
adjusted to account for the seasonal and temperature effects (FHWA, 2000).
The BELLS temperature prediction model was developed after work done by
Southgate (1959). Subsequently, several modifications were made to the BELLS model
123
that resulted in an improved model called BELLS2. The Seasonal Monitoring Program
(SMP) of the Long Term Pavement Performance (LTTP) program developed the most
comprehensive temperature and deflection data set ever to be assembled. The data was
used to develop a model that can be used to predict the temperature within an asphalt
layer from surface temperature data collected during routine deflection testing (FHWA,
2000). This model is the BELLS3 model and was used for this research project. Mid-
depth asphalt temperatures can be determined from Equation 4.1.
)5.13sin()(042.0)]5.15sin(83.1)1(621.0)(448.0[()25.1)(log()(892.095.0
1818 −+−+−+−⋅−++=
hrIRhrdayIRdIRTd Eqn. 4.1
where: Td = Pavement temperature at depth d, °C IR = Infrared surface temperature, °C log = base 10 logarithm d = depth at which mat temperature is to be predicted, mm 1-day = Average air temperature the day before testing
sin = Sine function on an 18-hr clock system, with 2π radians equal to one 18-hr cycle
hr18 = Time of day, in 24-hr clock system, but calculated using an 18-hr asphalt concrete (AC) temperature rise-and fall-time
Mid-depth asphalt temperatures were determined and input into the Evercalc program.
Evercalc then adjusts all the deflection measurements to a standard temperature of 25°C
(77°F).
4.4.1.2 DARWin
AASHTOWare DARWin v. 3.1.002 software was used by the MaineDOT to
backcalculate composite and subgrade moduli. This program is based on AASHTO
deflection analysis procedures. DARWin does not provide individual layer moduli, only
a composite modulus for asphalt and subbase layers and a modulus for the subgrade
124
layer. This was beneficial because the Prima 100 PFWD also provides composite
moduli, which allowed for a direct comparison. However, it was necessary to select
another program for backcalculation analysis to differentiate between moduli associated
with asphalt, subbase, and subgrade structural layers.
4.4.1.3 Evercalc
Evercalc 5.0, developed by the Washington State Department of Transportation
(WSDOT) was used for backcalculating FWD data to obtain individual layer moduli.
Evercalc is a pavement analysis computer program that estimates the “elastic” moduli of
pavement layers. Evercalc estimates the elastic modulus for each pavement layer,
determines the coefficients of stress sensitivity for unstabilized materials, stresses and
strains at various depths, and optionally normalizes asphalt concrete modulus to a
standard laboratory condition (temperature). Evercalc uses an iterative approach in
changing the moduli in a layered elastic solution to match theoretical and measured
deflections (WSDOT, 1999).
The Evercalc program uses WESLEA (provided by the Waterways Experiment
Station, U.S. Army Corps of Engineers) as the layered elastic solution to compute the
theoretical deflections and a modified Augmented Gauss-Newton algorithm for
optimizations. Basic assumptions of layered elastic theory include the following:
• Layers are infinitely long in the horizontal directions • Layers have uniform thickness • Bottom layer is semi infinite in the vertical direction • Layers are composed of homogeneous, isotropic, linearly elastic
materials, characterized by elastic modulus and Poisson’s ratio.
125
To begin the backcalculation process a general file must be created. The general
file allows for basic input parameters including but not limited to the following: loading
plate radius, number of sensors, sensor spacing, number of layers, and Poisson’s ratio.
Additional input parameters are shown in Figure 4.4. Locations are selected or
deselected for analysis and layer thicknesses are input. Options pertaining to the
treatment of multiple drops at several different load levels are available. This is shown in
Figure 4.5. Once the raw data file has been converted, a deflection data file is created.
This is shown in Figure 4.6. The deflection data may be modified if the measurements do
not follow a descending pattern moving radially outward from the center sensor.
Figure 4.4 Evercalc 5.0 general file data entry screen (WSDOT, 2001).
126
Figure 4.5 Evercalc 5.0 raw FWD data conversion screen (WSDOT, 2001).
Figure 4.6 Evercalc 5.0 FWD deflection data file screen (WSDOT, 2001).
127
4.4.2 Asphalt Surfaced Roads
Asphalt surfaced roads used for tracking seasonal stiffness variations include:
Kennebec Road, Buffalo Road, Stinson Lake Road, Knapp Airport Parking Lot, Witter
Farm Road, Route 126, and Route 1A. Backcalculation procedures are described in the
previous sections.
Prima 100 PFWD composite modulus, and for sites where it is available, FWD
asphalt, subbase, subgrade, and composite modulus and Loadman PFWD composite
modulus values are plotted versus date in Figures 4.7 through 4.20. In general, the
moduli are high when the pavement section is frozen and during the early part of the
period when section is partially thawed. At some field sites there are significant
differences in moduli from nearby test locations and from one week to the next. This
behavior was especially evident at Kennebec Road (Figures 4.7 and 4.8), Witter Farm
Road (Figures 4.12 through 4.14), and Route 126 (Figure 4.15).
As air temperatures fluctuate, the rate at which heat is added and/or removed from
the roadway changes. As a result, thawing and re-freezing may occur causing the
modulus to also change. This behavior was noticeable at Kennebec Road (Figure 4.3 and
4.4), Witter Farm Road (Figure 4.8 through 4.10), and Route 1A (Figure 4.14 through
4.16). At these sites partial thawing occurred at or near March 16, 2004 before the return
of freezing temperatures. A distinct increase in modulus occurred at each site on
approximately March 24, 2004. All three sites are located within a 32 km (20 mi) radius
of Bangor, Maine.
The composite moduli generally decreased as thawing progressed. It was
anticipated that a distinct minimum modulus would be reached near the end of the
128
thawing period followed by increasing modulus due to drainage of excess water in the
subbase and subgrade soils. This behavior was observed at the Buffalo Road (Figure
4.9), Knapp Airport Parking Lot (Figure 4.11), and to a lesser extent, the Stinson Lake
Road (Figure 4.10) test sites. All three sites reached distinct minimum values at or near
the end of March. At the remaining sites, the composite modulus at the end of the
thawing period was approximately equal to, or in some cases greater than, the value
measured in late June or early July. To better illustrate this behavior, the average PFWD
and FWD composite moduli at the end of the thawing period and in mid to late June are
summarized in Table 4.6.
Table 4.6 Summary of PFWD and FWD composite moduli at the end of thawing and during recovery periods.
Modulus on Date of Complete Thaw Modulus in mid to late June
Field Test Site
Test Section
Date of Complete
Thaw
Average Prima 100
PFWD Composite Modulus
(MPa)
Average FWD
Composite Modulus
(MPa)
Date of Final
Reading
Average Prima 100
PFWD Composite Modulus
(MPa)
Average FWD
Composite Modulus
(MPa)
Control 4/20 524 446 434 476 2 3/25 999 751 291 252
Witter Farm Road 1 4/16 466 347
6/28 348 351
3 4/14 427 533 280 468 Route 126 8 4/7 362 410 6/29 263 446 D-1 3/30 636 506 469 586 D-2 3/16 438 413 308 403 Route
1A E-3 3/23 457 474
6/28 319 551
Stinson Lake Road
1 4/19 279 NA 6/9 179 NA
NA – no composite modulus for spring thaw field test sites outside of Maine.
129
Backcalculated layer moduli are used as the basis for PFWD comparison. FWD derived
moduli indicate that Kennebec Road, Buffalo Road, and Knapp Airport Parking Lot
exhibited some degree of thaw weakening and recovery. Moduli derived from Prima 100
PFWD measurements also follow similar trends. Prima 100 PFWD moduli also followed
similar trends to FWD moduli at test sites where no thaw weakening occurred. Three of
these test sites: Witter Farm Road, Route 126, and Route 1A were all fully reconstructed
within the last ten years. Each test site was constructed with non frost susceptible
materials and as a result, none experienced thaw weakening as shown in plots
corresponding to those test sites. Based on these observations, the Prima 100 PFWD can
be used as a tool to aid in determining both when to apply and remove load restrictions.
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
7/17/04
Date
0
100
200
300
400
500
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD subgrade modulusAverage FWD composite modulus
Thawed
Figure 4.7 Stiffness variation at Kennebec Road (Section 1), Hampden/Dixmont, Maine.
130
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
7/17/04
Date
0
100
200
300
400
500
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD composite modulusAverage FWD subgrade modulus
Thawed
Figure 4.8 Stiffness variation at Kennebec Road (Section 2), Hampden/Dixmont, Maine.
3/13/043/31/04
4/18/045/6/04
5/24/046/11/04
Date
0
200
400
600
800
1000
1200
Mod
ulus
(MPa
)
Prima 100 PFWD - TP#1Prima 100 PFWD - TP#2Prima 100 PFWD - TP#3Prima 100 PFWD - TP#4Prima 100 PFWD - TP#5Prima 100 PFWD - TP#6Prima 100 PFWD - TP#7Prima 100 PFWD - TP#8Prima 100 PFWD - TP#9Prima 100 PFWD - TP#10
Avg. Loadman PFWDAvg. FWD Subbase ModulusAvg. FWD Subgrade Modulus
ThawedPartially Frozen
Figure 4.9 Stiffness variation at Buffalo Road, Rumney, New Hampshire.
131
3/13/043/31/04
4/18/045/6/04
5/24/046/11/04
Date
0
1000
2000
3000
Mod
ulus
(MPa
)
Prima 100 PFWD - TP #1Prima 100 PFWD - TP #2Prima 100 PFWD - TP #3Prima 100 PFWD - TP #4Prima 100 PFWD - TP #5Prima 100 PFWD - TP #6Prima 100 PFWD - TP #7Prima 100 PFWD - TP #8Prima 100 PFWD - TP #9Prima 100 PFWD - TP #10
Avg. Loadman PFWDAvg. FWD Subbase ModulusAvg. FWD Subgrade Modulus
Partially Frozen Thawed
Figure 4.10 Stiffness variation at Stinson Lake Road, Rumney, New Hampshire.
3/13/043/31/04
4/18/045/6/04
5/24/046/11/04
Date
0
200
400
600
800
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6PFWD - TP#7PFWD - TP#8PFWD - TP#9PFWD - TP#10Average FWD Subbase ModulusAverage FWD Subgrade Modulus
Thawed
Figure 4.11 Stiffness variation at Knapp Airport Parking Lot, Berlin, Vermont.
132
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
1000
2000
3000
4000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6Average FWD subgrade modulusAverage FWD composite modulus
Thawed
Partially Frozen
Figure 4.12 Stiffness variation at Witter Farm Road (Control Section), Orono, Maine.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
1000
2000
3000
4000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6Average FWD subgrade modulusAverage FWD composite modulus
Thawed
Partially Frozen
Frozen
Figure 4.13 Stiffness variation at Witter Farm Road (Section 2), Orono, Maine.
133
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
1000
2000
3000
4000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6Average FWD subgrade modulusAverage FWD composite modulus
ThawedPartially FrozenFrozen
Figure 4.14 Stiffness variation at Witter Farm Road (Section 1), Orono, Maine.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
7/17/04
Date
0
400
800
1200
1600
2000
Mod
ulus
(MPa
)
PFWD - STA 1+652PFWD - STA 1+664PFWD - STA 1+676PFWD - STA 1+688Average FWD subgrade modulusAverage FWD composite modulus
Frozen
Partially Frozen
Thawed
Figure 4.15 Stiffness variation at Route 126 (Section 3), Monmouth/Litchfield, Maine.
134
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
7/17/04
Date
0
400
800
1200
1600
2000
Mod
ulus
(MPa
)
PFWD - STA 4+028PFWD - STA 4+036PFWD - STA 4+044PFWD - STA 4+052Average FWD subgrade modulusAverage FWD composite modulus
Frozen
Partially Frozen
Thawed
Figure 4.16 Stiffness variation at Route 126 (Section 8), Monmouth/Litchfield, Maine.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
7/17/04
Date
0
400
800
1200
1600
2000
Mod
ulus
(MPa
)
PFWD - STA 4+704PFWD - STA 4+708PFWD - STA 4+712PFWD - STA 4+716Average FWD subgrade modulusAverage FWD composite modulus
Thawed
Figure 4.17 Stiffness variation at Route 126 (Section 12), Monmouth/Litchfield, Maine.
135
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
500
1000
1500
2000
Mod
ulus
(MPa
)
PFWD - STA 255+50PFWD - STA 256+50PFWD - STA 257+50PFWD - STA 258+50PFWD - STA 259+00PFWD - STA 260+50Average FWD subgrade modulusAverage FWD composite modulus
Frozen
ThawedPartially Frozen
Figure 4.18 Stiffness variation at Route 1A (Section D-1), Frankfort/Winterport, Maine.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
500
1000
1500
2000
Mod
ulus
(MPa
)
PFWD - STA 262+00PFWD - STA 263+00PFWD - STA 264+00PFWD - STA 265+00PFWD - STA 266+00PFWD - STA 267+00Average FWD subgrade modulusAverage FWD composite modulus
Frozen Thawed
Partially Frozen
Figure 4.19 Stiffness variation at Route 1A (Section D-2), Frankfort/Winterport, Maine.
136
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
500
1000
1500
2000
Mod
ulus
(MPa
)
PFWD - STA 291+00PFWD - STA 292+00PFWD - STA 293+00PFWD - STA 294+00PFWD - STA 295+00Average FWD subgrade modulusAverage FWD composite modulus
Frozen ThawedPartially Frozen
Figure 4.20 Stiffness variation at Route 1A (Section D-3), Frankfort/Winterport, Maine.
4.4.3 Gravel Surfaced Roads
Gravel surfaced roads tested for seasonal stiffness variations include Lakeside
Landing Road (Glenburn, Maine), Crosstown Road (Berlin, Vermont), and the USFS
Parking Lot (Rumney, New Hampshire). Additional portable device measurements were
taken at the USFS Parking Lot during the spring of 2003, these results are also presented.
Prima 100 PFWD composite modulus, and for sites where it is available, FWD
composite, subbase, and subgrade moduli are plotted versus date in Figures 4.17 through
4.23. In general, the moduli are high when the section is frozen and during the early part
of the period when section is partially thawed. At some field sites there are significant
differences in moduli from nearby test locations and from one week to the next. This is
more apparent in gravel surfaced test sites compared to asphalt surfaced test sites.
137
The composite moduli generally decreased as thawing progressed. It was
anticipated that a distinct minimum modulus would be reached near the end of the
thawing period followed by increasing modulus due to drainage of excess water in the
base and subgrade soils. This behavior was more apparent in the gravel surfaced test
sites where environmental factors and material uniformity have an increased effect on
measured moduli. The composite modulus measured in late June was approximately
equal to, or in most cases greater than, the value measured during the thaw period.
Measurements taken at the USFS Parking Lot in the spring of 2003 do not show the same
trends. The first set of measurements was taken at the end of April after the thawing
period had occurred. Thus, Figure 4.25 only illustrates a small portion of the recovery
period. Similar to asphalt surfaced test sites, gravel surfaced sites also showed the
effects of re-freezing. This was noticeable at the Lakeside Landing Road, USFS Parking
Lot (2004), and Crosstown Road test sites.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
1000000
2000000
3000000
4000000
5000000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD subgrade modulusAverage FWD composite modulus
ThawedPartially FrozenFrozen
Figure 4.21 Stiffness variation at Lakeside Landing Road (Section 1), Glenburn, Maine.
138
4/18/045/6/04
5/24/046/11/04
6/29/047/17/04
Date
0
300
600
900
1200
Mod
ulus
(MPa
)PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD subgrade modulusAverage FWD composite modulus
ThawedPartially Frozen
Figure 4.22 Detailed stiffness variation at Lakeside Landing Road (Section 1), Glenburn, Maine.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
400000
800000
1200000
1600000
2000000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD subgrade modulusAverage FWD composite modulus
ThawedPartially Frozen
Figure 4.23 Stiffness variation at Lakeside Landing Road (Section 2), Glenburn, Maine.
139
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
300
600
900
1200
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4Average FWD subgrade modulusAverage FWD composite modulus
ThawedPartially Frozen
Figure 4.24 Detailed stiffness variation at Lakeside Landing Road (Section 2), Glenburn, Maine.
4/19/035/7/03
5/25/036/12/03
6/30/037/18/03
Date
0
50
100
150
200
Mod
ulus
(MPa
)
Average Prima 100 PFWD ModulusAverage Loadman PFWD ModulusAverage Clegg Impact Hammer ModulusAverage Humboldt Soil Stiffness Gauge Modulus
Figure 4.25 2003 stiffness variation at USFS Parking Lot, Rumney, New Hampshire.
140
3/13/043/31/04
4/18/045/6/04
5/24/046/11/04
Date
0
400
800
1200
1600
2000M
odul
us (M
Pa)
Prima 100 PFWD - TP#3Prima 100 PFWD - TP#4Prima 100 PFWD - TP#5Prima 100 PFWD - TP#8Prima 100 PFWD - TP#9Prima 100 PFWD - TP#10Average Loadman PFWD
ThawedPartially Frozen
Figure 4.26 2004 stiffness variation at USFS Parking Lot, Rumney, New Hampshire.
3/31/044/18/04
5/6/045/24/04
6/11/04
Date
0
50
100
150
200
250
Mod
ulus
(MPa
)
Prima 100 PFWD - TP#3Prima 100 PFWD - TP#4Prima 100 PFWD - TP#5Prima 100 PFWD - TP#8Prima 100 PFWD - TP#9Prima 100 PFWD - TP#10Average Loadman PFWD
ThawedPartially Frozen
Figure 4.27 2004 detailed stiffness variation at USFS Parking Lot, Rumney, New Hampshire.
141
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/04
Date
0
2000
4000
6000
8000
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6PFWD - TP#7PFWD - TP#8PFWD - TP#9PFWD - TP#10Average FWD subbase modulus
Partially Frozen Thawed
Frozen
Figure 4.28 Stiffness variation at Crosstown Road, Berlin, Vermont.
3/31/044/18/04
5/6/045/24/04
6/11/04
Date
0
300
600
900
1200
Mod
ulus
(MPa
)
PFWD - TP#1PFWD - TP#2PFWD - TP#3PFWD - TP#4PFWD - TP#5PFWD - TP#6PFWD - TP#7PFWD - TP#8PFWD - TP#9PFWD - TP#10Average FWD subbase modulus
Thawed
Partially Frozen
Figure 4.29 Detailed stiffness variation at Crosstown Road, Berlin, Vermont.
142
4.5 COMPARISON OF PFWD AND FWD MODULI
Portable and traditional FWD derived moduli for the Route 11 and Route 167 test
sites are presented and assessed in this section. In addition, Prima 100 PFWD derived
composite moduli are compared to FWD derived composite and subbase moduli for both
asphalt and gravel surfaced test sites. Finally, portable and traditional FWD derived
impact stiffness moduli are compared for asphalt and gravel surfaced test sites.
4.5.1 Route 11 & Route 167 Field Test Sites
A single set of Prima 100 PFWD and FWD measurements were taken on Routes
11 in Wallagrass Plantation and Route 167 in Presque Isle/Fort Fairfield, Maine in May
2003 as part of an ongoing MaineDOT research project (Bouchedid and Humphrey,
2004). Composite modulus values derived from both PFWD and FWD as well as
subgrade moduli are plotted versus test location for each site. These results are presented
in Figures 4.30 through 4.37.
Prima 100 PFWD composite moduli follow a similar trend to that of the
traditional FWD at each test location at both test sites. Prima 100 PFWD composite
moduli are less than the composite moduli backcalculated from FWD data at all test
points. Prima 100 PFWD composite moduli are greater than subgrade moduli
backcalculated from FWD data at all test locations.
143
1 2 3 4 5Location
0
400
800
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.30 Modulus versus test location at Route 11 (Test Pit 1), Wallagrass Plantation, Maine.
1 2 3 4 5Location
0
400
800
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.31 Modulus versus test location at Route 11 (Test Pit 2), Wallagrass Plantation, Maine.
144
1 2 3Location
40
400
800
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.32 Modulus versus test location at Route 11 (Test Pit 3), Wallagrass Plantation, Maine.
1 2 3 4 5Location
0
400
800
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.33 Modulus versus test location at Route 11 (Test Pit 4), Wallagrass Plantation, Maine.
145
1 2 3 4 5Test Point
0
300
600
900
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.34 Modulus versus test location at Route 167 (Test Pit 1), Presque Isle/Fort Fairfield, Maine.
1 2 3Test Point
40
300
600
900
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.35 Modulus versus test location at Route 167 (Test Pit 2), Presque Isle/Fort Fairfield, Maine.
146
1 2 3 4 5Test Point
0
300
600
900
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.36 Modulus versus test location at Route 167 (Test Pit 3), Presque Isle/Fort Fairfield, Maine.
1 2 3 4 5Test Point
0
300
600
900
1200
Mod
ulus
(MPa
)
Prima 100 PFWD Composite ModulusFWD Composite ModulusFWD Subgrade Modulus
Figure 4.37 Modulus versus test location at Route 167 (Test Pit 4), Presque Isle/Fort Fairfield, Maine.
147
4.5.2 Composite Modulus
Composite moduli derived from traditional FWD measurements were supplied to
the researchers by MaineDOT for asphalt and gravel surfaced test sites located in the
State of Maine. For each site, backcalculated moduli are plotted against composite
moduli as measured with the Prima 100 PFWD in Figure 4.38 through 4.43. Regression
analyses yielded correlation coefficients ranging from 0.336 (Route 1A) to 0.950 (Witter
Farm Road). In general terms, correlation coefficients tended to increase as pavement
thickness decreased. To better illustrate this, separate plots were developed for sites with
different asphalt thicknesses. These are presented in Figures 4.44 through 4.46 and
Figure 4.43. Three test sites with asphalt thicknesses less than or equal to 127 mm (5 in.)
produced the best correlation with r2 = 0.873. Two test sites with an asphalt thickness of
152 mm (6 in.) followed with r2 = 0.559. However, when excluding moduli greater than
4000 MPa the correlation improves with r2 = 0.802. Route 1A served as the single test
site with a 180 mm (7 in.) asphalt thickness and produced the poorest correlation with r2
= 0.336. Data from all paved sites is presented in Figure 4.47. Regression analysis
yielded a correlation coefficient of 0.531, however, when excluding all moduli greater
than 4000 MPa (Figure 4.48), the correlation improved with r2 = 0.809. Results from the
Lakeside Landing Road test site are shown in Figure 4.49. Regression analysis yielded
an r2 of 0.446. Overall, a strong correlation exists between the Prima 100 PFWD
composite moduli and FWD derived composite moduli for asphalt surfaced roads. A
marginal correlation exists for gravel surfaced test sites.
Mean moduli for individual asphalt thicknesses are presented in Table 4.7. The
FWD and PFWD composite moduli are lower for the 178 mm (7 in.) asphalt thickness
148
than the 127 mm (5 in.) thickness. This is contrary to expectations, since thicker
pavements would be expected to yield higher composite moduli.
Table 4.7 FWD and PFWD mean composite moduli for different asphalt thicknesses.
Asphalt Thickness mm (in.)
Prima 100 PFWD Mean Composite Modulus
(MPa)
FWD Mean Composite Modulus
(MPa) 127 (5) 645 557 150 (6) 483 658 180 (7) 503 505
0 4000 8000 12000 16000Prima 100 PFWD Composite Modulus (MPa)
0
4000
8000
12000
16000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 1.354*X + 110.868Coef of determination, R-squared = 0.554
Figure 4.38 Comparison of FWD and PFWD composite moduli at Kennebec Road, Hampden/Dixmont, Maine.
149
0 500 1000 1500 2000 2500Prima 100 PFWD Composite Modulus (MPa)
0
500
1000
1500
2000
2500
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.699*X + 104.521Coef of determination, R-squared = 0.897
Figure 4.39 Comparison of FWD and PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine.
0 1000 2000 3000 4000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.807*X - 15.922Coef of determination, R-squared = 0.950
Figure 4.40 Comparison of FWD and PFWD composite moduli at Witter Farm Road, Orono, Maine.
150
0 200 400 600 800Prima 100 PFWD Composite Modulus (MPa)
0
200
400
600
800
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.947*X + 155.289Coef of determination, R-squared = 0.909
Figure 4.41 Comparison of FWD and PFWD composite moduli at Route 11, Wallagrass Plantation, Maine.
0 200 400 600 800 1000Prima 100 PFWD Composite Modulus (MPa)
0
200
400
600
800
1000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 1.493*X + 73.898Coef of determination, R-squared = 0.748
Figure 4.42 Comparison of FWD and PFWD composite moduli at Route 167, Presque Isle/Fort Fairfield, Maine.
151
0 200 400 600 800 1000 1200Prima 100 PFWD Composite Modulus (MPa)
0
200
400
600
800
1000
1200
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.421*X + 293.133Coef of determination, R-squared = 0.336
Figure 4.43 Comparison of FWD and PFWD composite moduli at Route 1A, Frankfort/Winterport, Maine.
0 1000 2000 3000 4000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.746*X + 76.39Coef of determination, R-squared = 0.873
Figure 4.44 Comparison of FWD and PFWD composite moduli for asphalt thicknesses ≤ 127 mm (5 in.).
152
0 4000 8000 12000 16000Prima 100 PFWD Composite Modulus (MPa)
0
4000
8000
12000
16000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 1.339*X + 11.151Coef of determination, R-squared = 0.559
Figure 4.45 Comparison of FWD and PFWD composite moduli for asphalt thicknesses equal to 152 mm (6 in.).
0 1000 2000 3000 4000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.747*X + 177.978Coef of determination, R-squared = 0.802
Figure 4.46 Comparison of FWD and PFWD composite moduli for asphalt thicknesses equal to 152 mm (6 in.) and moduli ≤ 4000 MPa.
153
0 4000 8000 12000 16000Prima 100 PFWD Composite Modulus (MPa)
0
4000
8000
12000
16000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 1.107*X - 35.7695Coef of determination, R-squared = 0.531
Figure 4.47 Comparison of FWD and PFWD composite moduli for all asphalt surfaced test sites.
0 1000 2000 3000 4000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.702*X + 147.172Coef of determination, R-squared = 0.809
Figure 4.48 Comparison of FWD and PFWD composite moduli for all asphalt surfaced test sites and moduli ≤ 4000 MPa.
154
0 200 400 600 800Prima 100 PFWD Composite Modulus (MPa)
0
200
400
600
800
FWD
Com
posit
e M
odul
us (M
Pa)
Equation Y = 0.390*X + 178.215Coef of determination, R-squared = 0.446
Figure 4.49 Comparison of FWD and PFWD composite moduli at Lakeside Landing Road, Glenburn, Maine.
4.5.3 Subbase Modulus
Subbase moduli were derived from traditional FWD data using Evercalc
backcalculation software. For each site, subbase moduli are plotted against composite
moduli as measured with the Prima 100 PFWD. These are presented in Figure 4.50
through 4.56. Regression analyses yielded correlation coefficients ranging from 0.163
(Route 1A) to 0.807 (Knapp Airport Parking Lot). In general, correlation coefficients
tended to decrease as pavement thickness increased. With thin asphalt layers, the
subbase modulus has an increased affect on PFWD composite moduli whereas with thick
asphalt layers, composite moduli are heavily influenced by the pavement layer. To better
illustrate this, separate plots were developed for sites with different asphalt thicknesses.
Five test sites with an asphalt thicknesses equal to 127 mm (5 in.) followed with r2 =
155
0.508. However, when excluding moduli greater than 5000 MPa, the correlation
improves with r2 = 0.693. This is shown in Figures 4.57 and 4.58. One test site (Route
126) with a 150 mm (6 in.) asphalt thickness, shown in Figure 4.55, produced the best
correlation with r2 = 0.698. Route 1A served as the single test site with a 180 mm (7 in.)
asphalt thickness, shown in Figure 4.56, and produced the poorest correlation with r2 =
0.363. Data from all paved sites is presented in Figure 4.59. Regression analysis yielded
a correlation coefficient of 0.485, however, when excluding all moduli greater than 5000
MPa (Figure 4.60), the correlation improved with r2 = 0.654. The results from the
Crosstown Road, gravel surfaced test site are presented in Figure 4.61. Composite and
subbase moduli compared marginally with r2 = 0.327. In general terms, PFWD
composite moduli had a reasonable correlation with FWD derived subbase moduli,
suggesting that the Prima 100 composite moduli are influenced at least in part by the
subbase layer.
0 5000 10000 15000 20000 25000Prima 100 PFWD Composite Modulus (MPa)
0
5000
10000
15000
20000
25000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 2.804*X - 364.938Coef of determination, R-squared = 0.651
Figure 4.50 Comparison of FWD subbase moduli and PFWD composite moduli at Kennebec Road, Hampden/Dixmont, Maine.
156
0 300 600 900 1200Prima 100 PFWD Composite Modulus (MPa)
0
300
600
900
1200
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.283*X - 36.704Coef of determination, R-squared = 0.593
Figure 4.51 Comparison of FWD subbase moduli and PFWD composite moduli at Stinson Lake Road, Rumney, New Hampshire.
0 100 200 300 400Prima 100 PFWD Composite Modulus (MPa)
0
100
200
300
400
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.298*X - 8.141Coef of determination, R-squared = 0.473
Figure 4.52 Comparison of FWD subbase moduli and PFWD composite moduli at Buffalo Road, Rumney, New Hampshire.
157
0 100 200 300 400Prima 100 PFWD Composite Modulus (MPa)
0
100
200
300
400
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.531*X - 26.222Coef of determination, R-squared = 0.807
Figure 4.53 Comparison of FWD subbase moduli and PFWD composite moduli at Knapp Airport Parking Lot, Berlin, Vermont.
0 1000 2000 3000 4000 5000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
5000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.917*X - 321.123Coef of determination, R-squared = 0.776
Figure 4.54 Comparison of FWD subbase moduli and PFWD composite moduli at Witter Farm Road, Orono, Maine.
158
0 400 800 1200 1600 2000Prima 100 PFWD Composite Modulus (MPa)
0
400
800
1200
1600
2000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.495*X - 4.325Coef of determination, R-squared = 0.698
Figure 4.55 Comparison of FWD subbase moduli and PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine.
0 400 800 1200Prima 100 PFWD Composite Modulus (MPa)
0
400
800
1200
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.163*X + 62.135Coef of determination, R-squared = 0.163
Figure 4.56 Comparison of FWD subbase moduli versus PFWD composite moduli at
Route 1A, Frankfort/Winterport, Maine.
159
0 5000 10000 15000 20000 25000Prima 100 PFWD Composite Modulus (MPa)
0
5000
10000
15000
20000
25000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 1.962*X - 569.247Coef of determination, R-squared = 0.508
Figure 4.57 Comparison of FWD subbase moduli and PFWD composite moduli for asphalt thicknesses ≤ 127 mm (5 in.).
0 1000 2000 3000 4000 5000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
5000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.734*X - 127.238Coef of determination, R-squared = 0.693
Figure 4.58 Comparison of FWD subbase moduli and PFWD composite moduli for asphalt thickness ≤ 127 mm (5 in.) and moduli ≤ 5000 MPa.
160
0 5000 10000 15000 20000 25000Prima 100 PFWD Composite Modulus (MPa)
0
5000
10000
15000
20000
25000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 1.869*X - 592.698Coef of determination, R-squared = 0.485
Figure 4.59 Comparison of FWD subbase moduli and PFWD composite moduli for all asphalt surfaced test sites.
0 1000 2000 3000 4000 5000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
5000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.689*X - 125.315Coef of determination, R-squared = 0.654
Figure 4.60 Comparison of FWD subbase moduli and PFWD composite moduli for all asphalt surfaced test sites and moduli ≤ 5000 MPa.
161
0 2000 4000 6000 8000 10000Prima 100 PFWD Composite Modulus (MPa)
0
2000
4000
6000
8000
10000
FWD
Sub
base
Mod
ulus
(MPa
)
Equation Y = 0.609*X + 362.733Coef of determination, R-squared = 0.327
Figure 4.61 Comparison of FWD subbase moduli and PFWD composite moduli at Crosstown Road, Berlin, Vermont.
4.5.4 Impact Stiffness Modulus
Impact Stiffness Modulus (ISM) is defined as the ratio of the applied load to the
deflection of the center sensor, or geophone, as seen by Equation 4.2.
0D
PISM = Eqn. 4.2
where: ISM = Impact Stiffness Modulus P = Applied load, kN (kip) D0 = Surface deflection at the center of the test load, μm (mils).
ISM values were determined for both the traditional and portable FWD. The advantage
of ISM over composite modulus is that use of a modulus backcalculation program is
avoided. The results for asphalt and gravel surfaced test sites are presented in the
following sections.
162
Plots of FWD derived ISM are compared to corresponding PFWD values and are
presented in Figures 4.62 through 4.68. The correlation tends to improve as pavement
thickness decreases. To better illustrate this, separate plots were developed for sites with
different asphalt thicknesses. These are shown in Figures 4.69 through 4.70, and Figure
4.68. All asphalt surfaced test sites are represented in Figure 4.71. This trend is similar
to that exhibited by composite and subbase moduli. This discussion is provided in
Section 4.5.2 and 4.5.3. Traditional and portable FWD derived ISM are compared for
three gravel surfaced test sites in Figures 4.72 through 4.74. Regression analyses yielded
correlation coefficients ranging from 0.638 (Lakeside Landing Road) to 0.914
(Crosstown Road). All gravel surfaced sites were combined and are shown in Figure
4.75. As a result, the PFWD and FWD derived ISM moduli appear to be equally
effective indicators of section stiffness for thin asphalt surfaced and gravel surfaced sites.
0 0.4 0.8 1.2 1.6 2Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.4
0.8
1.2
1.6
2
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.998*X + 0.065Coef of determination, R-squared = 0.765
Figure 4.62 Comparison of FWD and PFWD ISM at Kennebec Road, Hampden/Dixmont, Maine.
163
0 0.1 0.2 0.3 0.4Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.1
0.2
0.3
0.4
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 1.285*X - 0.017Coef of determination, R-squared = 0.905
Figure 4.63 Comparison of FWD and PFWD ISM at Stinson Lake Road, Rumney, New Hampshire.
0 0.02 0.04 0.06 0.08 0.1Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.02
0.04
0.06
0.08
0.1
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.868*X + 0.015Coef of determination, R-squared = 0.7597
Figure 4.64 Comparison of FWD and PFWD ISM at Buffalo Road, Rumney, New Hampshire.
164
0 0.02 0.04 0.06 0.08 0.1Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.02
0.04
0.06
0.08
0.1
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.716*X + 0.015Coef of determination, R-squared = 0.755
Figure 4.65 Comparison of FWD and PFWD ISM at Knapp Airport Parking Lot, Berlin, Vermont.
0 0.2 0.4 0.6 0.8 1Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.2
0.4
0.6
0.8
1
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.904*X - 0.005Coef of determination, R-squared = 0.937
Figure 4.66 Comparison of FWD and PFWD ISM at Witter Farm Road, Orono, Maine.
165
0 0.1 0.2 0.3 0.4 0.5Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.1
0.2
0.3
0.4
0.5
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 1.025*X + 0.011Coef of determination, R-squared = 0.899
Figure 4.67 Comparison of FWD and PFWD ISM at Route 126, Monmouth/Litchfield, Maine.
0 0.1 0.2 0.3Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.1
0.2
0.3
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.593*X + 0.064Coef of determination, R-squared = 0.488
Figure 4.68 Comparison of FWD and PFWD ISM at Route 1A, Frankfort/Winterport, Maine.
166
0 0.2 0.4 0.6 0.8 1Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.2
0.4
0.6
0.8
1
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.897*X + 0.007Coef of determination, R-squared = 0.922
Figure 4.69 Comparison of FWD and PFWD ISM for test sites with asphalt thicknesses ≤ 127 mm (5 in.).
0 0.4 0.8 1.2 1.6 2Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.4
0.8
1.2
1.6
2
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 1.0098*X + 0.0346Coef of determination, R-squared = 0.768
Figure 4.70 Comparison of FWD and PFWD ISM for test sites with asphalt thicknesses equal to 152 mm (6 in.).
167
0 0.4 0.8 1.2 1.6 2Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.4
0.8
1.2
1.6
2
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.959*X + 0.013Coef of determination, R-squared = 0.797
Figure 4.71 Comparison of FWD and PFWD ISM for all asphalt surfaced test sites.
0 0.2 0.4 0.6 0.8
Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.2
0.4
0.6
0.8
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.27*X + 0.032Coef of determination, R-squared = 0.638
Figure 4.72 Comparison of FWD and PFWD ISM at Lakeside Landing Road, Glenburn, Maine.
168
0 0.4 0.8 1.2 1.6Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.4
0.8
1.2
1.6
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.622*X + 0.033Coef of determination, R-squared = 0.914
Figure 4.73 Comparison of FWD and PFWD ISM at Crosstown Road, Berlin, Vermont.
0 0.005 0.01 0.015 0.02 0.025 0.03Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.005
0.01
0.015
0.02
0.025
0.03
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.417*X + 0.0007Coef of determination, R-squared = 0.810
Figure 4.74 Comparison of FWD and PFWD ISM at USFS Parking Lot, Rumney, New Hampshire.
169
0 0.4 0.8 1.2 1.6Prima 100 PFWD Impact Stiffness Modulus (kN/μm)
0
0.4
0.8
1.2
1.6
FWD
Impa
ct S
tiffn
ess M
odul
us (k
N/μ
m)
Equation Y = 0.6196*X + 0.021Coef of determination, R-squared = 0.902
Figure 4.75 Comparison of FWD and PFWD ISM for all gravel surfaced test sites.
4.6 COMPARISON TO OTHER PORTABLE DEVICES
The traditional FWD was used as the basis for comparison for the Prima 100
PFWD, Loadman PFWD, Humboldt Soil Stiffness Gauge (SSG), and Clegg Impact
Hammer. Loadman PFWD measurements were taken at Stinson Lake Road, Buffalo
Road, and the USFS Parking Lot (2003 and 2004). SSG measurements were taken at the
USFS Parking Lot during the spring of 2003. Clegg Impact Hammer measurements were
taken at the USFS Parking Lot site during the spring of 2003 and 2004. A description of
each of the portable devices is provided in Section 2.2.1. The Clegg Impact Hammer,
Soil Stiffness Gauge, and the Prima 100 PFWD do not share a common variable with the
FWD that can be compared. As a result, comparisons are only made between the Prima
170
100 and Loadman PFWDs. Comparisons between the devices and the FWD were
developed and are presented separately in the following section.
Loadman and Prima 100 PFWD composite moduli are compared to FWD derived
subbase moduli for two asphalt surfaced test sites in Rumney, New Hampshire. Best fit
lines correlating the two devices with FWD subbase moduli are shown in Figure 4.76. In
addition, Loadman composite moduli are plotted versus date and presented in Figures
4.5, 4.6, 4.21, 4.22, and 4.23. Correlations developed for asphalt surfaced test sites
indicate that for a given FWD derived subbase modulus, the Loadman PFWD provides a
composite modulus which is greater than the corresponding value provided by the Prima
100. The Prima 100 PFWD correlates better to FWD derived subbase moduli (r2 =
0.552) than composite moduli obtained from the Loadman PFWD (r2 = 0.245). The
Loadman PFWD uses a smaller loading plate diameter and drop weight. As a result,
higher moduli are obtained because only the upper most pavement section is influenced.
Correlations developed for gravel surfaced test sites indicate that for a given Prima 100
PFWD composite modulus, the corresponding Loadman PFWD modulus is lower due to
the shallow depth of influence.
Based on these results, it appears that the Prima 100 PFWD is a better tool to aid
in tracking seasonal stiffness variations. In addition, the Prima 100 is more versatile than
the Loadman as discussed in Section 2.2.2.1 and 2.2.2.2. Prima 100 PFWD input
parameters allow the user to differentiate between asphalt, gravel, and other materials by
selecting appropriate stress distribution factors and Poisson’s ratio. However, this
recommendation is based on results from only two asphalt surfaced test sites (Buffalo
Road and Stinson Lake Road).
171
0 1000 2000 3000 4000PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Sub
base
Mod
ulus
(MPa
)
Prima 100 PFWDLoadman PFWDPrima 100 PFWDLoadman PFWD
Equation Y = 0.246*X - 9.655Coef of determination, R-squared = 0.552
Equation Y = 0.058*X + 29.794Coef of determination, R-squared = 0.245
Figure 4.76 Comparison of FWD derived subbase moduli to Loadman and Prima 100 PFWD composite moduli on asphalt surfaced test sites.
4.7 EVALUATION OF FIELD TESTING TECHNIQUES
A number of different testing techniques were developed and implemented. The
effect of drop weight, loading plate diameter, and drop height was investigated. In
addition, multiple measurements were taken at each test location in order to examine the
extent to which moduli change with subsequent drops. Lastly, additional geophones were
used at each test site to investigate their usefulness. The influence of these variations on
testing techniques are presented in the following sections.
4.7.1 Loading Plate Diameter and Drop Weight
The Keros Prima 100 PFWD was purchased in March 2003. At that time, the
device came standard with 100 and 300 mm (4 and 12 in.) diameter loading plates and
172
one 10 kg (22 lb) drop weight. A 200 mm (8 in.) diameter loading plate and two 5 kg (11
lb) weights were also purchased. A preliminary study on an asphalt surfaced parking lot
was undertaken to investigate the differences in moduli derived from using different
combinations of drop weight and loading plate diameter. Measurements were taken at
three different locations. The details of the findings are presented in Figure 4.77, and are
discussed below.
In general, the 20 kg (44 lb) drop weight produced the lowest moduli. Moduli
were independent of loading plate diameter. The 15 kg (33 lb) weight produced moduli
that were greater than those obtained with the 20 kg (44 lb.) and also did not vary
significantly with loading plate diameter. The highest moduli were obtained using the 10
kg (22 lb) drop weight. Moduli decreased with increasing loading plate diameter.
0 1 2 3 4Test Location
0
200
400
600
800
1000
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
10 kg Drop Weight15 kg Drop Weight20 kg Drop Weight
100 mm Loading Plate200 mm Loading Plate
300 mm Loading Plate
Figure 4.77 Effect of drop weight and loading plate diameter on Prima 100 PFWD composite moduli.
173
Reducing the drop weight to 10 kg (22 lb) results in a significantly higher modulus and as
noted above, plate diameter has a larger effect. Small plate diameter and drop weight
influence only the upper portions of the pavement section and thus the deflection
responses are dominated by the stiffer pavement layer, producing a larger composite
modulus. When plate diameter and drop weight are increased, depth of influence is
increased and the stiffness of both the subbase and asphalt layers are reflected in the
composite modulus, resulting in a lower value. During the spring thaw period, it is
desirable to measure the stiffness of the subbase layer since it, not the asphalt layer, is
more likely to undergo thaw weakening and recovery. As a result, it is recommended
that the largest loading plate and drop weight be used in order to maximize the influence
on the subbase layer.
4.7.2 Drop Height
Drop heights ranging from 10 to 850 mm (0.4 to 33.5 in.) may be used with the
Prima 100 PFWD. Measurements were taken at three different drop heights at each test
location throughout the monitoring period. Drop heights used were approximately equal
to 850, 630, and 420 mm (33.5, 24.8, and 16.5 in.). Plots with best fit lines comparing
moduli derived from different drop heights with FWD composite moduli are presented in
Figures 4.78 through 4.80. In general, reduced drop heights produce moduli that are
slightly less than moduli derived from using the full (850 mm) drop height. This trend is
evident regardless of asphalt thickness, however, the differences tend to decrease with
increasing asphalt thickness as illustrated in Table 4.8. Decreasing drop height reduces
the depth of influence. When asphalt thickness increases and drop height is reduced
simultaneously, measured moduli are heavily influenced by the stiffness of the asphalt
174
layer. For the purpose of monitoring seasonal stiffness variations, it is desirable to
influence the greatest depth possible. As a result, it is recommended that the 850 mm
(33.5 in.) drop height be used.
Table 4.8 Summary of the effects of reduced drop height on PFWD composite modulus for different asphalt thicknesses.
Test Site
Asphalt Thickness mm (in.)
FWD Composite Modulus
(MPa)
Drop Height(mm)
Prima 100 PFWD
Composite Modulus
(MPa)
Percent Difference
(%)
850 2498 24.9 630 2444 22.2
Witter Farm Road
127 (5) 2000 420 2271 13.6 850 960 4.0 630 909 9.1
Route 126 152 (6) 1000
420 889 11.1 850 6430 114.3 630 5651 88.4
Route 1A 178 (7) 3000
420 5663 88.8
0 1000 2000 3000 4000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
FWD
Com
posit
e M
odul
us (M
Pa)
850 mm Drop Height630 mm Drop Height420 mm Drop Height
850 mm drop height
630 mm drop height
420 mm drop height
Figure 4.78 Effect of drop height on PFWD composite moduli at Witter Farm Road, Orono, Maine.
175
0 400 800 1200 1600 2000Prima 100 PFWD Composite Modulus (MPa)
0
400
800
1200
1600
2000
FWD
Com
posit
e M
odul
us (M
Pa)
850 mm Drop Height630 mm Drop Height420 mm Drop Height
850 mm drop height
420 mm drop height
630 mm drop height
Figure 4.79 Effect of drop height on PFWD composite moduli at Route 126, Monmouth/Litchfield, Maine.
0 500 1000 1500Prima 100 PFWD Composite Modulus (MPa)
0
500
1000
1500
FWD
Com
posit
e M
odul
us (M
Pa)
850 mm Drop Height630 mm Drop Height420 mm Drop Height
850 mm Drop Height630 mm Drop Height420 mm Drop Height
Figure 4.80 Effect of drop height on PFWD composite moduli at Route 1A, Frankfort/Winterport, Maine.
176
4.7.3 Moduli Derived from Additional Geophones
The Keros Prima 100 PFWD comes standard with one geophone. Two additional
deflection sensors were purchased. Three deflection sensors were used at each site to
observe differences in moduli derived from measurements taken from each of the
geophones. Spacing of the sensors is as follows (as measured from the center of the
loading plate): 0, 207, and 407 mm (0, 8, and 16 in.). Prima 100 PFWD software
determines moduli from the measurements of just one geophone at a time by using the
Boussinesq equations described in Section 2.2.2. Thus, when three geophones are used,
three calculations for modulus are performed, yielding three estimates of composite
modulus. In addition, unsuccessful attempts were made to use backcalculation software
intended for a conventional FWD using all three geophone measurements in order to
derive moduli for individual layers. Future research should be conducted to develop
backcalculation software for the Prima 100 PFWD that uses all three sensors to estimate
layer moduli. Results for asphalt and gravel surfaced roads are presented in Figures 4.81
through 4.83. Unless software can be developed to incorporate deflections from
additional geophones into a more common backcalculation routine, additional geophones
do not provide meaningful information. Based on the quality of the correlations between
FWD derived moduli and PFWD moduli determined from the center geophone, it is
recommended that only one geophone be used.
177
Table 4.9 Summary of the effects of reduced drop height on PFWD composite moduli for different asphalt thicknesses.
Test Site Asphalt
Thickness mm (in.)
FWD Composite Modulus
(MPa)
RadialOffset (mm)
Prima 100 PFWD Composite Modulus
(MPa)
Percent Difference
(%)
0 23057 131 207 97586 8759
Kennebec Road 127 (5) 10000
407 119162 1092 0 5247 5
207 27608 452 Route 126
152 (6)
5000 407 24513 390 0 4054 103
207 17346 767 Route 1A
178 (7) 2000 407 20397 920
0 10000 20000 30000 40000Prima 100 PFWD Composite Modulus (MPa)
0
10000
20000
30000
40000
FWD
Com
posit
e M
odul
us (M
Pa)
0 mm radial offset207 mm radial offset407 mm radial offset
0 mm radial offset207 mm radial offset
407 mm radial offset
Figure 4.81 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Kennebec Road, Hampden/Dixmont, Maine.
178
0 2000 4000 6000 8000 10000Prima 100 PFWD Composite Modulus (MPa)
0
2000
4000
6000
8000
10000
FWD
Com
posit
e M
odul
us (M
Pa)
0 mm radial offset207 mm radial offset407 mm radial offset
0 mm radial offset
207 mm radial offset
407 mm radial offset
Figure 4.82 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Route 126, Monmouth/Litchfield, Maine.
0 1000 2000 3000 4000 5000 6000Prima 100 PFWD Composite Modulus (MPa)
0
1000
2000
3000
4000
5000
6000
FWD
Com
posit
e M
odul
us (M
Pa)
0 mm radial offset207 mm radial offset407 mm radial offset
0 mm radial offset
407 mm radial offset
207 mm radial offset
Figure 4.83 Comparison of FWD composite moduli to PFWD composite moduli derived from different geophones at Route 1A, Frankfort/Winterport, Maine.
179
4.7.4 Multiple Measurements at Each Test Location
Previous researches noted differences in moduli when taking multiple
measurements at the same test location. Many researchers performed multiple
measurements at each test location and suggested disregarding one or more of the initial
readings, using the remaining measurements to determine a representative value. Six
Prima 100 PFWD measurements were taken at each of three different drop heights, at
each test location. In all cases, the first reading was neglected and the average of the
remaining five was used for analysis and comparison. Observations by the researchers
during field testing indicated that for the majority of points tested at all field sites; the
first measurement was less than subsequent measurements. The difference between the
first measurement and subsequent measurements is depicted in Figure 4.84 and Figure
4.85.
1 2 3 4 5Drop Number
6180
200
220
240
260
280
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
STA 1+652STA 1+664STA 1+676STA 1+688
Figure 4.84 Effect of consecutive drops on composite modulus on May 12, 2004 at Route 126 (Section 3), Monmouth/Litchfield, Maine.
180
1 2 3 4 5Drop Number
6250
300
350
400
450
500
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
STA 4+704STA 4+708STA 4+712STA 4+716
Figure 4.85 Effect of consecutive drops on composite modulus values on April 22, 2004 at Route 126 (Section 12), Monmouth/Litchfield, Maine.
In order to quantify this phenomenon, all measurements taken at Route 126 in the
towns of Monmouth and Litchfield, Maine were analyzed. The first, second, and third
measurements were compared to the average of the remaining measurements. These
results are shown in Table 4.10. The observations confirm those made by other
researchers discussed in Chapter 2. In any case, no matter the total number of drops
selected for each test location and the material being tested, at least the first measurement
should be ignored. The representative value should be determined from an average of the
remainder of measurements taken at that test location.
181
Table 4.10 Comparison of the first, second, and third measurements with successive measurements at Route 126, Monmouth/Litchfield, Maine.
Average Value of
First Measurement (MPa)
Average of Five Remaining Measurements
(MPa)
Percent Difference
(%) 388 428 9.4
Average Value of Second Measurement
(MPa)
Average of Four Remaining Measurements
(MPa)
Percent Difference
(%) 426 430 0.95
Average Value of Third Measurement
(MPa)
Average of Three Remaining Measurements
(MPa)
Percent Difference
(%) 428 429 0.35
4.8 RECOMMENDATIONS
There are few straight forward procedures to aid in determining the need for
weight restrictions, the magnitude of the restriction, and when to place and remove the
restriction from paved and unpaved, low volume roads. The basis of the methods for
placing and removing load restrictions include observing the pavement structure for signs
of distress, measuring surface deflections, and more recently, predicting thaw from air
temperature data. For a load restriction policy to be implemented successfully it must be
as simple as possible, yet include the most important factors common to the greatest
number of roadway miles (Ovik, et al., 2000). In the following, we recommend
procedures for using a PFWD to determine if a road should be posted for weight
restriction and then a procedure to determine the duration of the restriction.
182
4.8.1 Factors that Affect Need for Seasonal Load Restrictions
As discussed by Rutherford, et al. (1985) many factors exist that should be
considered when determining whether seasonal load restrictions are necessary at a
particular location. These factors are listed below.
1. Pavements with surface deflections 45 to 50% higher during spring thaw than
summer.
2. Pavements with frost susceptible subbase and subgrade material.
3. Pavement with subgrade soils classified as ML, MH, CL, and CH.
4. Roads which have historically exhibited deterioration during the spring thaw
period.
5. Pavements in which distress has been observed (fatigue cracking and rutting).
The procedure that is recommended in the next section for use of the Prima 100 PFWD
should be applied with due consideration of the factors listed above.
A procedure for using the PFWD to place and remove load restrictions is
presented in the following section. The procedure is then applied to the field sites
monitored as part of this study.
4.8.2 Field Testing Techniques
Field testing techniques for monitoring seasonal stiffness variation in paved and
unpaved low volume roads using the Prima 100 PFWD have been developed. The
recommendations are based on the experiences of the researchers in using the Prima 100
PFWD as discussed in previous sections and the techniques developed by previous
researchers (Rutherford, et al., 1985; Van Deusen 1998; Ovik, et al., 2000). The
183
procedure relies on comparing composite moduli during the spring thaw to fully
recovered values measured during the summer and fall. Thus, the underlying premise is
that composite modulus is the primary factor controlling damage to the road section.
The researchers selected 80% of the fully recovered composite modulus as the
trigger value for application and removal of load restrictions. The selection of 80% is
arbitrary since the amount of damage that would occur at the reduced modulus depends
on individual pavement sections, allowable vehicle weight, and traffic levels.
Assessment of these factors was beyond the scope of this study. Individual transportation
agencies should examine these issues in light of the amount of damage that is acceptable
to the road during the spring thaw period and the consequences to the regional economy
that are created by weight restrictions.
The procedure recommended by the researchers is outlined in the following steps.
It can be used to determine when to apply and remove load restrictions. In addition, it
can be used as a screening procedure to identify roads that do not require posting.
1. For each road to be monitored, identify critical sections of the road that are most
susceptible to spring thaw damage based on pervious performance, soil type,
access to ground water, or other factors. Within each critical section, select four
test points. Test points should span the inside and outside wheelpaths in both
travel lanes, if present. The location of the points should be marked so that the
same locations can be tested on each test day.
2. Setup the Prima 100 PFWD with the 850 (33.5 in.) drop height, 20 kg (44 lb) drop
weight, and 300 mm (12 in.) diameter loading plate.
184
3. Setup the Personal Digital Assistant (PDA) based recording software using the
input parameters presented in Table 4.11.
Table 4.11 Prima 100 PFWD input parameters.
Setup Menu Item
Input Parameter
Asphalt SurfacedTest Sites
Gravel Surfaced Test Sites
Pretrig time (ms) 10* Pulsebase (%) 24* Trigger
Trig Level (kN) 0.90* View Sample Time (ms) 60*
Load Plate Radius (mm) 150 Number of sensors 1 Mechanical
D(1) offset (cm) 0 Poisson’s Ratio 0.35** 0.35** Formula
Stress Distribution 2.0 2.67 * - default values ** - Huang (2004)
4. Establish moduli for each test point that are representative of the fully recovered
period by taking readings at each test point during the summer and early fall.
Readings should be taken on days that correspond to periods that are relatively
dry. A reading at an individual test point is the average of drops 2 through 6. The
results from drop 1 are discarded. It is recommended that readings be taken on
four days spanning the summer and early fall. Average the four daily readings at
each point to obtain the fully recovered composite modulus for that point.
Finally, average the recovered composite modulus from each test point to obtain
the recovered composite modulus for the section. Multiply this value by 80% to
obtain the trigger value for load restriction application/removal.
5. Using the same test points and testing techniques that were used to establish the
baseline values, take periodic readings at the start of the spring thaw. During the
185
critical thawing period, it may be necessary to take readings daily. Taking
readings in the afternoon is preferred to avoid the influence of possible refreezing
during the previous night. A reading at an individual test point is the average of
drops 2 through 6. The results from drop 1 are discarded. Apply the load
restriction when the average of the composite moduli at the test points in the
section drops below 80% of the baseline (recovered) values.
6. Continue to take periodic readings, at least weekly. Once the average of the
composite moduli at the test points in the section readings exceed 80% of the
baseline (recovered) values for two consecutive sets of measurements the load
restriction may be removed.
7. Sites where the moduli remain above 80% of the recovered value are potential
candidates for roads that do not require posting.
4.8.3 Application of Procedure to Field Sites
Of the ten test sites monitored during the spring of 2004, four showed distinct
minimum composite moduli during the thawing period before increasing into the
recovery period. Asphalt surfaced test sites were Buffalo Road and Knapp Airport
Parking Lot. Gravel surfaced sites were the USFS Parking Lot and Crosstown Road.
Application of the criteria described in the previous section for placing and removing
load restriction is applied to spring thaw test sites and is shown in Figures 4.86 through
4.89. The procedure was also used with available results from the FWD. A summary of
the dates for placing and removing the restrictions are summarized in Table 4.11. In
186
general, the posting and removal dates determined by the PFWD and FWD agree within
one week.
It should be noted that the recovered composite modulus used in the application
of the procedure recommended in Section 4.8.2 were based on a single reading date in
June rather than the average of four reading dates in late summer and early fall. Thus,
the interpretation of the duration of the load restriction may have been somewhat
different had the latter readings been available. Moreover, readings were taken weekly,
whereas the composite modulus can experience a dramatic reduction over this period as
illustrated in Figures 4.88 and 4.89. This shows the importance of taken readings more
frequently during the thawing period.
Table 4.12 Summary of load restrictions for spring thaw field test sites.
Restriction Posting
Date
Date of Minimum Modulus
Restriction Removal
Date
Date of Final
Reading Test Site
PFWD FWD PFWD FWD PFWD FWD PFWD FWD Buffalo Road 3/25 3/18 4/1 3/25 4/15 4/8 6/9 6/9
USFS Parking Lot 4/1 NA 4/1 NA 5/14 NA 6/9 NA Knapp Airport
Parking Lot NA NA 3/26 4/2 5/14 NA 6/10 6/10
Crosstown Road 4/2 4/2 4/2 4/9 5/14 5/14 6/10 6/10 NA – not available – could not be determined from available data.
This research found that roads that have undergone full depth reconstruction with
125 mm (5 in.) or more of pavement supported by 600 mm (24 in.) of non-frost
susceptible base (Witter Farm Road, Route 1A, and Route 126) did not experience a
seasonal reduction in the composite resilient modulus, as shown in Figures 4.12 through
4.20, and thus do not require seasonal weight restrictions.
187
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/04
Date
100
200
300
Mod
ulus
(MPa
)
Average Prima 100 PFWD composite modulusAverage FWD subbase modulus
80% of final measurement (PFWD)
80% of final measurement (FWD)
FWDLoad Restriction
PFWDLoad Restriction
Figure 4.86 Buffalo Road, Rumney, New Hampshire.
2/24/043/13/04
3/31/044/18/04
5/6/045/24/04
6/11/046/29/04
Date
0
100
200
300
Mod
ulus
(MPa
)
Average Prima 100 PFWD composite modulusAverage FWD subbase modulus
80% of final measurement (PFWD)
PFW
DR
emov
e L
oad
Res
tric
tion
80% of final measurement (FWD)
Figure 4.87 Knapp Airport Parking Lot, Berlin, Vermont.
188
3/6/043/24/04
4/11/044/29/04
5/17/046/4/04
6/22/04
Date
0
100400500600700800900
100011001200
Ave
rage
Pri
ma
100
PFW
D C
ompo
site
Mod
ulus
(MPa
)
80% of final measurement
PFWDLoad Restriction
Figure 4.88 USFS Parking Lot, Rumney, New Hampshire.
3/6/043/24/04
4/11/044/29/04
5/17/046/4/04
Date
0
5003000
3500
4000
4500
5000
Mod
ulus
(MPa
)
Average Prima 100 PFWD Composite ModulusAverage FWD Subbase Modulus
PFWD & FWDLoad Restriction
80% of final measurement (FWD)
80% of final measurement (PFWD)
Figure 4.89 Crosstown Road, Berlin, Vermont.
189
4.9 SUMMARY
The results of monitoring seasonal stiffness variations in paved and unpaved,
seasonally posted, low volume roads were presented and assessed in this chapter. For
sites where the traditional FWD indicated a decrease in composite modulus during the
spring thaw, the Prima 100 PFWD followed similar trends. Similarly, at sites where the
FWD did not show a decrease in composite modulus, neither did the PFWD. Based on
this result, the Prima 100 PFWD would equally effective as a traditional FWD in
determining when to place and remove load restrictions.
Prima 100 PFWD composite moduli were compared to composite moduli derived
from the traditional FWD. Regression analyses comparing composite moduli from both
devices yielded correlation coefficients for individual sites ranging from 0.336 (Route
1A) to 0.950 (Witter Farm Road). In general, PFWD composite moduli were slightly less
than FWD composite moduli. Correlation coefficients tended to increase with
decreasing pavement thickness. When combining all the results for paved roads, a strong
correlation exists between Prima 100 PFWD and FWD derived composite. Ignoring
errant FWD moduli greater than 4000 MPa ( 41,770 tsf), the regression coefficient for all
the data combined was 0.809. Looking at subsets of the data confirmed that the
regression coefficient increased at the pavement thickness decreased. This suggests that
the PFWD could be used in lieu of an FWD for determination of composite modulus.
Regression analyses comparing composite and subbase moduli yielded correlation
coefficients ranging from 0.163 (Route 1A) to 0.807 (Knapp Airport Parking Lot).
Again, correlation coefficients tended to increase as pavement thickness decreased. The
190
PFWD had a reasonable correlation with FWD derived subbase moduli, suggesting that
the Prima 100 composite moduli are at influenced in part by the subbase layer.
Impact Stiffness Modulus (ISM) is the ratio of applied load to the deflection
measured by the center sensor (geophone). ISM values were determined for both the
traditional and portable FWD. Regression analyses for paved test sites yielded
correlation coefficients ranging from 0.488 (Route 1A) to 0.937 (Witter Farm Road).
Correlations coefficients increased with decreasing asphalt thickness. Regression
analyses for gravel test sites yielded correlation coefficients ranging from 0.638
(Lakeside Landing Road) to 0.914 (Crosstown Road). PFWD and FWD derived ISM
appear to equally effective indicators of section stiffness for sites with asphalt pavements
less than about 150 mm (6 in.) and gravel surfaced sites.
Loadman and Prima 100 PFWD derived composite moduli were compared to
subbase moduli as determined from the traditional FWD for two asphalt surfaced test
sites. This was done to determine which device would serve as a better tool for
evaluating seasonal stiffness variations. The Prima 100 produced larger composite
moduli than the Loadman and correlated better with the FWD derived moduli, producing
an r2 = 0.552. In contrast, the Loadman PFWD produced an r2 = 0.245. It was
recommended that the Prima 100 PFWD would serve as a better tool to aid in tracking
seasonal stiffness variations.
Several testing techniques were used to observe their influence on composite
moduli. The effect of drop weight, loading plate diameter, and drop height were
investigated. Additionally, multiple measurements were taken at each test point with
additional geophones. The lowest drop weight (10 kg) resulted in significantly higher
191
moduli, most likely because results for the lowest weight were primarily influenced by
the stiffer pavement layer. Loading plate diameter had little effect. The largest drop
weight and loading plate diameter are recommended for spring thaw monitoring.
Three different drop heights were used. Reduced drop heights produce moduli
that are slightly less than moduli derived from using the full drop height. Increased
asphalt thickness reduces the difference between moduli obtained from different drop
heights. It is recommended that the full drop height be used.
Three different geophones were used for testing. The geophones were located at
distances of 0, 207, and 407 mm (0, 8, and 16 in.) from the center of the drop plate.
Deflection measurements from each geophone were used to make three separate
calculations of modulus. The geophones located 207 and 407 mm (8 and 16 in.) from the
center produced moduli that are unrealistically large. It is recommended that only the
center geophone be used for testing. Future research should focus on developing
backcalculation software for the Prima 100 that would enable measurements from all
geophones to be simultaneously incorporated into a backcalculation routine.
Multiple PFWD measurements were taken at each test location to investigate how
composite moduli change with successive drops. The first drop differed from the average
of the remaining five measurements by nearly 10%. Whereas, the second drop differed
from the average of the remaining four measurements by approximately 1%. It was
recommended that the first measurement be neglected when determining the composite
modulus for a particular test location.
Finally, recommendations were made on how the Prima 100 PFWD could be used
to determine when spring load restrictions should be placed and removed, as well as,
192
roads where spring load restrictions were unnecessary. The core of the recommendations
are that the load restrictions are placed and removed once the composite moduli
measured with the PFWD reach 80% of the fully recovered baseline value measured
during the summer and early fall.
193
CHAPTER 5
COMPACTION CONTROL
5.1 INTRODUCTION
This chapter presents the analysis and results of the field and laboratory
evaluation of the Portable Falling Weight Deflectometer (PFWD) as an alternative to
traditional compaction control devices. The objective of this portion of the research
project was to establish a procedure for using the PFWD for compaction control. As part
of this effort, the relationship between PFWD composite moduli and percent compaction
for soil types representative of New England base and subbase aggregates was explored.
PFWD and Nuclear Moisture Density Meter (NDM) measurements were taken at
five field test sites during the summer and fall of 2003. In addition, laboratory tests were
completed on five different samples during the summer of 2004. The primary purpose of
the laboratory tests was to provide complementary results to those collected in field tests.
However, laboratory work was completed under more carefully controlled conditions.
Target dry densities for laboratory work were 90%, 95%, and 100% of the maximum dry
density (AASHTO T 180). The effect of water content was investigated at approximately
95% of the maximum dry density with target water contents equal to optimum and ± 3%
of the optimum water content (OWC). A more detailed description of laboratory testing
procedures may be found in Section 3.5. Comparative side by side tests of PFWDs by
multiple manufacturers were completed. Tests were done to investigate repeatability,
accuracy, and susceptibility to operator technique.
194
This chapter is organized as follows. Water content and density measurements
are compared and verified. This is followed by the effect of percent compaction and
water content on modulus (stiffness) observed in both field and laboratory tests.
Statistical relationships are discussed. Comparisons between portable devices are made
and correlations were developed to compare performance. The effect of operator
technique is discussed. Lastly, recommendations are made for utilizing the Prima 100
PFWD as a tool to monitor compaction control.
5.2 IN-PLACE WATER CONTENT AND DRY DENSITY
Laboratory tests were performed on five soil types representative of New England
base and subbase aggregates. These materials include: one crushed material, one
construction sand, three base/subbase aggregates. Classification and laboratory
compaction (AASTHO T 180) results for laboratory samples are summarized in Table
5.1. Gradation and moisture density curves for each sample may be found in Appendixes
A and B, respectively.
For the laboratory tests, each soil sample was compacted in the test box initially at
a low density. Measurements were taken and samples were then compacted to a higher
density and the measurements repeated. In total, 29 combinations of water content and
density were tested in the laboratory. For each trial, nuclear density gage (NDM)
measurements were made at five locations. NDM density measurements at the 203 mm
(8 in.) depth are used in this report unless noted otherwise. The average in-place dry
density, percent compaction, water content, and water content relative to optimum from
each trial as determined by the NDM is summarized in Table 5.2. For most of the trials,
195
the actual percent compaction and water content relative to optimum deviated somewhat
from the target values. Nonetheless, the range of values was sufficient to explore
relationships with PFWD composite modulus.
Table 5.1 Summary of laboratory samples.
Material Description Grain Size Characteristics Moisture Density Characteristics2
Type Name AASHTO Classification
Percent Gravel1
(%)
Percent Sand1
(%)
Percent Fines1
(%)
γd(max)Mg/m3
(lb/ft3)
wopt (%)
Crushed Material
Conn. crushed gravel
A-1-a 66.3 28.5 5.2 2.31 (144) 7.4
Construction Sand
N.H. sand A-1-b 24.6 72.6 2.8 2.06
(128) 10.6
N.H gravel A-1-b 45.0 52.1 2.9 2.05
(128) 9.2
OJF gravel A-1-b 34.2 61.3 4.5 2.00
(125) 11.2 Base/
Subbase Aggregate
Wardwell gravel A-1-a 51.9 42.1 6.0 2.10
(131) 5.1 1 – based on ASTM D 422. 2 – based on AASHTO T 180.
Water content values obtained from the NDM were compared to oven dried test
results in order to verify the accuracy of the NDM measurements. The NDM results were
taken to be the average of five readings, one at each test location, taken at a 203 mm (8
in.) depth. Oven dried water contents are compared to NDM water contents in Figure
5.1. In general, water contents determined from the NDM were greater than oven dried
samples. Oven dried samples produce true moisture content by removing all water
present. The NDM measures hydrogen present in the material, which typically is in the
form of water. If the material contains naturally occurring hydrogen or bound hydrogen,
the NDM will measure the moisture falsely high in many cases (Troxler, 2004). The
comparison produced a reasonable correlation with r2 = 0.549. As a result, water
196
contents determined from the NDM were used as the basis for analysis and comparison in
the remainder of this chapter.
Table 5.2 Summary of laboratory measurements.
Material Target Test
Average Dry Density
Mg/m3 (lb/ft3)
Average Percent
Compaction (%)
Average Water
Content (%)
Average Water
Content Relative to Optimum
(%) 90%, wopt 2.07 (129) 90 5.8 -1.6 95%, wopt 2.26 (141) 98 6.0 -1.4 100%, wopt 2.37 (148) 103 6.5 -0.9
95%, +3%wopt 2.32 (145) 100 5.5 -1.9
Con
nect
icut
cr
ushe
d gr
avel
95%, -3%wopt 2.15 (134) 93 4.3 -3.1 90%, wopt 1.78 (111) 86 8.2 -2.4
1.86 (116) 91 8.2 -2.4 1.91 (119) 92 8.9 -1.7 95%, wopt
1.91 (119) 94 9.0 -1.6 1.94 (121) 94 8.8 -1.8 95%, +3%wopt 1.87 (117) 91 11.7 +1.1
New
H
amps
hire
sa
nd
95%, -3%wopt 1.92 (120) 94 8.0 -2.6 90%, wopt 1.97 (123) 96 11.6 +2.4 95%, wopt 1.95 (122) 95 10.9 +1.7 100%, wopt 2.02 (126) 98 7.8 -1.4
95%, +3%wopt 2.02 (126) 99 9.5 +0.3 New
H
amps
hire
gr
avel
95%, -3%wopt 1.89 (118) 92 7.7 -1.5 90%, wopt 2.02 (126) 98 11.6 +0.4 95%, wopt 1.94 (121) 97 10.9 -0.3 100%, wopt 2.02 (126) 101 10.6 -0.6
95%, +3%wopt 2.02 (126) 101 11.5 +0.3
OJF
gr
avel
95%, -3%wopt 1.89 (118) 94 7.7 -3.5 90%, wopt 1.94 (121) 92 6.0 +0.9 95%, wopt 2.05 (128) 98 6.2 +1.1 100%, wopt 2.13 (133) 101 6.7 +1.6
1.91 (119) 91 8.1 +3.0 1.95 (122) 93 11.1 +6.0 95%, +3%wopt
1.89 (118) 90 15.3 +10.2
War
dwel
l gr
avel
95%, -3%wopt 2.03 (127) 97 4.0 -1.1
197
0 4 8 12 16Nuclear Density Gauge Water Content (%)
0
4
8
12
16
Ove
n D
ried
Wat
er C
onte
nt (%
)
Connecticut crushed gravelN.H. sandN.H. gravelOJF gravelWardwell gravel
Equation Y = 0.791*X + 0.640Coef of determination, R-squared = 0.549
Figure 5.1 Comparison of oven dried and NDM water contents.
In addition, the percent compaction as determined from sand cone tests was
compared to the percent compaction determined from the NDM. Like the water content
results, the percent compaction results from the NDM were taken to be the average of
five readings, one at each test location, taken at a 203 mm (8 in.) depth. The comparison
is shown in Figure 5.2. When the data points from all the projects were included, there
was essentially no correlation. However, it is difficult to perform accurate sand cone
tests in crushed gravel. If the data from the Connecticut crushed gravel are ignored, there
is a general trend of increasing percent compaction from the sand cone and NDM. The
sand cone predicted percent compactions that were greater than the NDM, and many
results were in the range of 100% to 123%. The upper end of the sand cone percent
compactions are unreasonable. NDM measurements were also taken at depths of 0, 51,
102, and 152 mm (0, 2, 4, and 6 in.), all of which exhibited a similar comparison with the
198
sand cone results. Given the unreasonably high percent compactions resulting from some
of the sand cone results, it was concluded that the NDM results were more reliable. In
the balance of this chapter, the percent compaction determined by the NDM was used as
the basis for comparison of the PFWD results.
75 100 125NDM Percent Compaction (%)
75
100
125
Sand
Con
e Pe
rcen
t Com
pact
ion
(%)
Connecticut crushed gravelN.H. sandN.H. gravelOJF gravelWardwell gravel
Equation Y = 0.066*X + 99.875Coef of determination, R-squared = 0.0006
Figure 5.2 Comparison of percent compaction determined from sand cone and NDM tests.
Furthermore, measurements were made at field sites located in Maine, New
Hampshire, and Connecticut. The field component included tests on two subgrades, one
construction sand product, two aggregates, and one reclaimed stabilized base product.
Classification and laboratory compaction results (AASTHO T 180) are summarized in
Table 5.3. Gradation and moisture density curves for each sample may be found in
Appendixes A and B, respectively. Overall, the water contents measured at the field sites
were significantly lower than those in the laboratory tests.
199
Table 5.3 Summary of field samples.
Material Description Grain Size Characteristics Moisture DensityCharacteristics2
Type Name AASHTO Classification
PercentGravel1
(%)
PercentSand1
(%)
Percent Fines1
(%)
γd(max)Mg/m3
(lb/ft3) wopt (%)
CPR A-1-a 34.9 54.1 11.0 2.05 (128) 5.5 Subgrade
I-84 A-1-b 28.5 57.4 14.1 NA NA
I-84 A-1-a 66.3 28.5 5.2 2.31 (144) 7.4
Route 25 A-1-b 36.8 59.4 3.8 1.92 (120) 12.3 Aggregate
Route 26 52.0 46.0 2.0 1.99 (124) 12.0
Construction Sand Route 25 A-1-b 36.8 60.4 2.8 2.17
(135) 11.4
CPR NA NA NA NA NA NA Reclaimed Stabilized Base Route 201 NA NA NA NA NA NA NA – not available 1 - based on ASTM D 422. 2 - based on AASHTO T 180.
5.3 FACTORS AFFECTING COMPOSITE MODULUS
The relationship of PFWD composite modulus as determined by the Prima 100
with percent compaction, water content, grain size distribution, and particle shape is
explored in this section. Each is discussed separately below. In addition, multiple
variable linear regression analysis is used to investigate the combined role of percent
compaction and water content on composite modulus. Portable testing devices are
compared and the effect of operator technique is discussed later in Section 5.4.
200
5.3.1 Effect of Percent Compaction on Composite Modulus
5.3.1.1 Laboratory Test Results
The Prima 100 PFWD composite modulus at each test location for each of the
varying percent compaction and water content trials is shown as bar graphs in Figures 5.3
through 5.7. The water content relative to OWC and the percent compaction shown in
the legend of these figures is the average of the values measured at the five test points for
the trial. For a given test location, the bar graphs are ordered from low to high water
content. The full dataset is shown in tabular form in Appendix C.
0 1 2 3 4 5 6Test Location
0
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
93% compaction, -3.1%wopt100% compaction, -1.9%wopt90% compaction, -1.6%wopt98% compaction, -1.4%wopt103% compaction, -0.9%wopt
Figure 5.3 Effect of percent compaction on composite modulus, Connecticut crushed gravel.
201
0 1 2 3 4 5 6Test Location
0
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
94% compaction, -2.6%wopt
86% compaction, -2.4%wopt
91% compaction, -2.4%wopt
94% compaction, -1.8%wopt
92% compaction, -1.7%wopt
94% compaction, -1.6%wopt
91% compaction, +1.1%wopt
Figure 5.4 Effect of percent compaction on composite modulus, New Hampshire sand.
0 1 2 3 4 5 60
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
92% compaction, -1.5%wopt
98% compaction, -1.4%wopt
99% compaction, +0.3%wopt
95% compaction, +1.7%wopt
96% compaction, +2.4%wopt
Figure 5.5 Effect of percent compaction on composite modulus, New Hampshire gravel.
202
0 1 2 3 4 5 6Test Location
0
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
94% compaction, -3.5%wopt
101% compaction, -0.6%wopt
97% compaction, -0.3%wopt
101% compaction, +0.3%wopt
98% compaction, +0.4%wopt
Figure 5.6 Effect of percent compaction on composite modulus, OJF gravel.
0 1 2 3 4 5 6Test Location
0
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
97% compaction, -1.1%wopt
92% compaction, +0.9%wopt
98% compaction, +1.1%wopt
101% compaction, +1.6%wopt
91% compaction, +3.0%wopt
93% compaction, +6.0%wopt
90% compaction, +10.2%wopt
Figure 5.7 Effect of percent compaction on composite modulus, Wardwell gravel.
203
Plots of composite modulus versus percent compaction as measured at each test
location were made. Single variable linear regression analyses (Neter, et al., 1982) were
used to determine the best fit line for composite modulus as a function of percent
compaction. The coefficient of simple determination (r2) was used to measure the degree
to which the variation of the dependent variable, in this case composite modulus, could
be explained by a linear relation with an independent variable, in this case percent
compaction (Walpole and Myers, 1978). For example, an r2 of 0.9 would indicated that
90% of the variation in composite modulus was explained by a linear relation with
percent compaction.
The plots of composite modulus versus percent compaction are shown in Figures
5.8 through 5.14. All data points regardless of water content are included. For four out
of the five materials tested, there was a general trend of increasing composite modulus
with increasing percent compaction. The r2 for these four materials ranged from 0.027 to
0.531. In contrast, the New Hampshire gravel exhibited the opposite trend with
decreasing composite modulus with increasing percent compaction and an r2 of 0.147.
Combining all laboratory samples produced an r2 of 0.069. This is shown in Figure 5.13.
However, when the two samples with the poorest correlation (New Hampshire gravel and
OJF gravel), the correlation improves with r2 = 0.312. The regression equations and r2
are summarized in Table 5.4. With the exception of the New Hampshire sand, the
regression coefficients were less then 0.5 indicating poor correlation. One reason for the
poor correlation could be that water content has an important influence on composite
modulus and this was not accounted for in Figures 5.8 through 5.14 or the regression
results in Table 5.4.
204
Table 5.4 Summary of the correlations between percent compaction and composite modulus for laboratory samples.
Sample Regression Equation
Coefficient of Simple Determination (r2)
Connecticut crushed gravel y = 2.524x – 100.731 0.122
New Hampshire sand y = 10.389x – 847.035 0.531
New Hampshire gravel y = -2.725x + 357.106 0.147
OJF gravel y = 0.663x + 3.899 0.027 Wardwell gravel y = 6.096x – 484.889 0.419
All Samples Combined y = 2.409x – 128.259 0.069
All Samples Combined (w/o NHG and OJF) y = 5.437 – 400.906 0.312
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 2.524*X - 100.731Coef of determination, R-squared = 0.122
Figure 5.8 Comparison of percent compaction and composite modulus, Connecticut crushed gravel.
205
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 10.389*X - 847.035Coef of determination, R-squared = 0.531
Figure 5.9 Comparison of percent compaction and composite modulus, New Hampshire sand.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = -2.725*X + 357.106Coef of determination, R-squared = 0.147
Figure 5.10 Comparison of percent compaction and composite modulus, New Hampshire gravel.
206
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 0.663*X + 3.899Coef of determination, R-squared = 0.027
Figure 5.11 Comparison of percent compaction and composite modulus, OJF gravel.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 6.096*X - 484.889Coef of determination, R-squared = 0.419
Figure 5.12 Comparison of percent compaction and composite modulus, Wardwell gravel.
207
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandNew Hampshire GravelOJF GravelWardwell Gravel
Equation Y = 2.409*X - 128.259Coef of determination, R-squared = 0.069
Figure 5.13 Comparison of percent compaction and composite modulus for all laboratory samples.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandWardwell Gravel
Equation Y = 5.437*X - 400.906Coef of determination, R-squared = 0.312
Figure 5.14 Comparison of percent compaction and composite modulus for three laboratory samples.
208
In an attempt to minimize the effect of water content, all laboratory data was
divided into two groups: one for test points dry of the OWC, and one for test points wet
of the OWC. Plots of composite modulus versus percent compaction for the two groups
are shown in Figure 5.15 and 5.17. The correlation coefficients for the two groups are
0.045 and 0.211, respectively. Again, the two samples with the poorest correlation (New
Hampshire gravel and OJF gravel) were removed. The correlation improves with r2 of
0.350 and 0.455. These plots are shown in Figures 5.16 and 5.18. Thus, subdividing the
data into those dry and wet of OWC did little to improve the correlation between
composite modulus and percent compaction.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandNew Hampshire GravelOJF GravelWardwell Gravel
Equation Y = 1.806*X - 56.086Coef of determination, R-squared = 0.045
Figure 5.15 Comparison of percent compaction and composite modulus for laboratory tests with water contents dry of the OWC.
209
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandWardwell Gravel
Equation Y = 5.000*X - 343.735Coef of determination, R-squared = 0.350
Figure 5.16 Comparison of percent compaction and composite modulus for selected laboratory samples with water contents dry of the OWC.
70 80 90 100 110 120Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
New Hampshire SandNew Hampshire GravelOJF GravelWardwell Gravel
Equation Y = 3.495*X - 254.469Coef of determination, R-squared = 0.211
Figure 5.17 Comparison of percent compaction and composite modulus for laboratory tests with water contents wet of the OWC.
210
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
New Hampshire SandWardwell Gravel
Equation Y = 5.660*X - 450.691Coef of determination, R-squared = 0.455
Figure 5.18 Comparison of percent compaction and composite modulus for selected laboratory samples with water contents wet of the OWC.
5.3.1.2 Field Test Results
The main objective of field site testing was to provide complementary
measurements to those obtained under more carefully controlled conditions in the
laboratory. The test sites had undergone varying degrees of compaction ranging from
uncompacted base to base that had been well compacted with vibratory smooth drum
rollers. Field testing techniques are described in Section 3.4.2. Compaction and soil
property data is summarized in Section 5.2. Base aggregate tested at I-84 and Route 25
test sites were the same as those tested in the laboratory. Maximum dry density and
optimum water content were unavailable for the I-84 subgrade soils, so the moduli from
these tests were compared directly to dry density. NDM measurements at Route 26 and
Route 201 test sections were taken prior to calibration. As a result, the NDM
211
measurements were deemed unreliable. Instead the change in composite modulus over
time is considered.
Prima 100 PFWD composite moduli are compared to percent compaction as
determined from the NDM for the Connecticut crushed gravel, New Hampshire sand, and
New Hampshire gravel. This is shown in Figures 5.19 through 5.23. The aggregate was
significantly dry of optimum (average of -4.4% for Connecticut crushed gravel, and
-9.0% for both New Hampshire sand and gravel). Results from the field test sites
indicate that as the degree of compaction increases, composite modulus also increases,
mirroring the observations made in the laboratory. The r2 and regression equations are
summarized in Table 5.5. The r2 for the Connecticut crushed gravel and New Hampshire
sand were greater than 0.5 indicating a reasonable correlation. The low r2 for the New
Hampshire gravel may be due in part to the small range of percent compactions for these
results. Results from these three sites were combined into a single plot as shown in
Figure 5.24. This resulted in an r2 of 0.818, indicating a reasonable degree of correlation.
The regression equation and r2 for the combined result is also shown in Table 5.5. The
regression equation suggests that for aggregate compacted at water contents at least 4.4%
drier than OWC, 100% of ASSHTO T180 compaction corresponds to a composite
modulus of 154 MPa. Examining the results for select laboratory tests in Figure 5.14
shows that 100% of AASHTO T 180 corresponds to a composite modulus of 143 MPa,
which is similar to the field result.
212
Table 5.5 Summary of the correlations between percent compaction and composite modulus for field samples.
Field Test
Site Material Regression Equation
Coefficient of Simple Determination (r2)
Crushed Gravel base y = 4.989x – 359.924 0.647 I-84 Subgrade* y = 0.953x – 33.463 0.014 Sand base y = 4.225x – 270.545 0.544 Route 25 Gravel base y = 1.598x + 21.572 0.008
CPR Subgrade y = 8.193x – 548.367 0.313 Three base materials combined
NA y = 5.75x – 420.736 0.818
NA – not applicable * - regression analysis results compare dry density and composite modulus
Reclaimed stabilized base products tested at the Route 201 and Commercial
Paving & Recycling field sites were monitored on multiple dates to examine the increase
in composite modulus over time. Results indicate an increase in composite modulus over
time for all stations monitored. This is shown in Figures 5.25 and 5.26. The additional
measurements were taken after completion of paving at both sites and both had been
opened to vehicle traffic. Some of the increase at the Route 201 test site could simply be
the result of increasing the pavement thickness since thicker pavement sections would
likely produce greater composite moduli. This suggests that the PFWD could be used to
monitor the time-dependent increases in composite modulus of asphalt stabilized base
materials.
213
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 4.989*X - 359.924Coef of determination, R-squared = 0.647
Figure 5.19 Comparison of percent compaction and composite modulus of crushed gravel tested at I-84, Southington, Connecticut.
100 110 120 130 140Dry Density (lb/ft3)
0
50
100
150
200
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 0.953*X - 33.463Coef of determination, R-squared = 0.014
Figure 5.20 Comparison of dry density and composite modulus of subgrade tested at I-84, Southington, Connecticut.
214
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 4.225*X - 270.545Coef of determination, R-squared = 0.544
Figure 5.21 Comparison of percent compaction and composite modulus of construction sand tested at Route 25, Effingham/Freedom, New Hampshire.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 1.598*X + 21.572Coef of determination, R-squared = 0.008
Figure 5.22 Comparison of percent compaction and composite modulus of gravel tested at Route 25, Effingham/Freedom, New Hampshire.
215
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
400
500
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 8.193*X - 548.367Coef of determination, R-squared = 0.313
Figure 5.23 Comparison of percent compaction and composite modulus of subgrade tested at CPR, Scarborough, Maine.
70 80 90 100 110 120Percent Compaction (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Conn. Crushed GravelN.H. GravelN.H. Sand
Equation Y = 5.75*X - 420.736Coef of determination, R-squared = 0.818
Figure 5.24 Comparison of percent compaction and composite modulus for materials tested at Route 25 and I-84 field test sites.
216
STA 0+20
STA 0+40
STA 0+60
STA 0+80
STA 1+40
STA 1+60
STA 1+800
100
200
300
400
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
10/24/200311/6/200311/21/2003
Figure 5.25 Change in moduli with time at CPR test site, Scarborough, Maine.
1240+471240+50
1240+531240+56
1240+591241+53
1241+561241+59
1241+621241+65
1242+501242+53
1242+561242+59
1242+620
100
200
300
400
500
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
8/6/20038/11/2003
Figure 5.26 Change in moduli with time at Route 201 test site, The Forks, Maine.
217
5.3.2 Effect of Water Content on Composite Modulus
5.3.2.1 Laboratory Test Results
Examination of the bar graphs previously shown in Figures 5.3 through 5.7 shows
that there is a general trend that the composite moduli tends to decrease as water content
increases. This was examined for the individual data points in Figures 5.27 through 5.32.
The equation for the best fit straight line and the associated r2 is shown on these figures
and summarized in Table 5.6. Correlation coefficients ranged from 0.003 (Connecticut
crushed gravel) to 0.814 (Wardwell gravel). The generally low correlation coefficients
are due in part to the role that percent compaction plays in the composite modulus, which
is not accounted for in this analysis. The higher correlation coefficient for the Wardwell
gravel may be due to the high water content of some of the samples as illustrated in
Figures 5.33 and 5.34. Wet of optimum, increased water content would tend to produce
lower densities, which also would contribute to a lower composite modulus and a
stronger correlation.
Table 5.6 Summary of the correlations between water content and composite modulus for laboratory samples.
Sample Regression Equation
Coefficient of Simple Determination (r2)
Connecticut crushed gravel y = -2.309x + 139.147 0.003
New Hampshire sand y = -10.005x + 90.036 0.112
New Hampshire gravel y = -8.385x + 97.876 0.407
OJF gravel y = -4.235x + 65.794 0.297 Wardwell gravel y = -11.009x + 126.181 0.814
All Samples Combined Y = -7.621x + 100.703 0.285
218
-4 -2 0 2 4Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (7
.4%
)Equation Y = -2.309*X + 139.147Coef of determination, R-squared = 0.003
Figure 5.27 Comparison of water content and composite modulus, Connecticut crushed gravel.
-4 -2 0 2 4Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (1
0.6%
)
Equation Y = -10.005*X + 90.036Coef of determination, R-squared = 0.112
Figure 5.28 Comparison of water content and composite modulus, New Hampshire sand.
219
-4 -2 0 2 4 6Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (9
.2%
)
Equation Y = -8.385*X + 97.876Coef of determination, R-squared = 0.407
Figure 5.29 Comparison of water content and composite modulus, New Hampshire gravel.
-5 -4 -3 -2 -1 0 1 2 3 4 5Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (1
1.2
%)
Equation Y = -4.235*X + 65.794Coef of determination, R-squared = 0.297
Figure 5.30 Comparison of water content and composite modulus, OJF gravel.
220
-4 0 4 8 12Water Content Relative to Optimum (%)
160
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (5
.1%
)
Equation Y = -11.009*X + 126.181Coef of determination, R-squared = 0.814
Figure 5.31 Comparison of water content and composite modulus, Wardwell gravel.
-8 -4 0 4 8 12Water Content Relative to Optimum (%)
160
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandNew Hampshire GravelOJF GravelWardwell Gravel
Equation Y = -7.621*X + 100.703Coef of determination, R-squared = 0.285
Figure 5.32 Comparison of water content and composite modulus for all laboratory samples.
221
Figure 5.33 Prima 100 PFWD measurement on Wardwell gravel wet of optimum.
Figure 5.34 Prima 100 PFWD measurement on Wardwell gravel wet of optimum.
222
5.3.2.2 Field Test Results
Prima 100 PFWD composite moduli are compared to water content as determined
from the NDM in Figures 5.35 through 5.40. Maximum dry density and optimum water
content were unavailable for the I-84 subgrade soils, so the moduli from these tests were
compared directly to water content. Regression analysis results for all field sites are
shown in Table 5.7. With the exception of the subgrade material at the I-84 test site, all
other materials exhibited trends indicating increasing modulus with increasing water
content. This is contrary to expectations and opposite of what was seen in laboratory
results. However, when the three base materials are combined, the trend is reversed,
mirroring laboratory observations. This is shown in Figure 5.40. The coefficient of
simple determination ranged from 0.008 (Route 25 gravel) to 0.521 (Route 25 sand).
One possible explanation for the low coefficients is that the water contents of base
materials the at field sites were all dry of the OWC and, at individual sites, spanned a
narrow range.
Table 5.7 Summary of the correlations between water content and composite modulus for field samples.
Field Test Site Material Regression
Equation
Coefficient of Simple
Determination (r2)
Crushed Gravel base y = 30.580x + 218.660 0.226 I-84
Subgrade* y = -19.525x + 203.639 0.412 Sand base y= 15.179x + 250.263 0.521 Route 25 Gravel base y = 2.133x + 208.394 0.008
CPR Subgrade y = 55.604x + 244.918 0.402 Three base materials combined
NA y = -16.85x + 1.135 0.288
NA – not applicable. * - correlated with density and water content.
223
-8 -6 -4 -2 0 2Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posit
e M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (7
.4%
)
Equation Y = 30.580*X + 218.660Coef of determination, R-squared = 0.226
Figure 5.35 Comparison of water content and composite modulus of crushed gravel tested at I-84, Southington, Connecticut.
0 2 4 6 8Water Content (%)
100
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = -19.525*X + 203.639Coef of determination, R-squared = 0.412
Figure 5.36 Comparison of water content and composite modulus of subgrade tested at I-84, Southington, Connecticut.
224
-10 -8 -6 -4 -2 0 2Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (1
0.6%
)
Equation Y = 15.179*X + 250.263Coef of determination, R-squared = 0.521
Figure 5.37 Comparison of water content and composite modulus of sand tested at Route 25, Effingham/Freedom, New Hampshire.
-12 -10 -8 -6 -4 -2 0 2 4Water Content Relative to Optimum (%)
0
100
200
300
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Opt
imum
Wat
er C
onte
nt (9
.2%
)
Equation Y = 2.133*X + 208.394Coef of determination, R-squared = 0.008
Figure 5.38 Comparison of water content and composite modulus of gravel tested at Route 25, Effingham/Freedom, New Hampshire.
225
-1 0 1 2Water Content Relative to Optimum (%)
0
100
200
300
400
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Equation Y = 55.604*X + 244.918Coef of determination, R-squared = 0.402
Opt
imum
Wat
er C
onte
nt (5
.5%
)
Figure 5.39 Comparison of water content and composite modulus of subgrade tested at CPR, Scarborough, Maine.
-10 -8 -6 -4 -Water Content Relative to Optimum (%)
20
100
200
300
400
500
Prim
a 10
0 PF
WD
Com
posi
te M
odul
us (M
Pa)
Connecticut Crushed GravelNew Hampshire SandNew Hampshire Gravel
Equation Y = -16.85 * X + 1.135Coef of determination, R-squared = 0.288
Figure 5.40 Comparison of water content and composite modulus for all field test sites.
226
5.3.3 Multivariate Linear Regression
Multivariable linear regression analysis (Neter, et al., 1982) was used to
determine the best fit line for composite modulus as a function of percent compaction and
water content. All data points regardless of water content are included. The coefficient
of multiple determination (R2) was used to measure the degree to which the variation of
the dependent variable, in this case composite modulus, could be explained by a linear
relation with two independent variables, in this case percent compaction and water
content (Walpole and Myers, 1978). For example, an R2 of 0.9 would indicate that 90%
of the variation in composite modulus was explained by a linear relation with percent
compaction and water content.
Coefficient of multiple determination as well as the regression equations are
provided for laboratory and field results in Tables 5.8 and 5.9, respectively. The R2 for
the laboratory materials ranged from 0.141 (Connecticut crushed gravel) to 0.867
(Wardwell gravel). Combining all laboratory samples produced an R2 of 0.326.
However, when the two samples with the poorest correlation between composite modulus
and percent compaction (New Hampshire Gravel and OJF Gravel) are removed from the
analysis, the correlation improves with R2 = 0.624. These R2 are significantly higher than
when only percent compaction is considered (R2 = 0.069 for all samples and R2 = 0.312
when NH gravel and OJF gravel are removed as, shown in Table 5.4), illustrating the
importance of considering water content in the empirical equation to predict composite
modulus. The R2 for the field materials ranged from 0.001 (Route 25 gravel) to 0.679 (I-
84 crushed gravel). Combining the field results for the three sites where base materials
were tested produced an R2 of 0.823 indicating a reasonably strong correlation between
227
Table 5.8 Summary of multivariate linear regression analyses on laboratory samples.
Sample Regression Equation
Coefficient of Multiple Determination (R2) for composite modulus as a function
of percent compaction and water content relative to
optimum Connecticut
crushed gravel Mc = -118.22 + 2.596(PC) -
5.946(RWC) 0.141
New Hampshire sand
Mc = -832.27 + 10.083(PC) – 8.153(RWC) 0.609
New Hampshire gravel Mc = 300.4 – 2.113(PC) – 7.341(RWC) 0.457
OJF gravel Mc = -99.275 + 1.672(PC) – 5.247(RWC) 0.451
Wardwell gravel Mc = -133.14 + 2.682(PC) – 9.268(RWC) 0.867
All Samples Combined
Mc = -77.989 + 1.878(PC) – 7.296(RWC) 0.326
Combined (CT, NHS, and
Wardwell)
Mc = -332.91 + 4.720(PC) – 7.658(RWC) 0.624
Mc = composite modulus; PC = percent compaction; RWC = relative water content. Table 5.9 Summary of multivariate linear regression analyses on field tests on
granular base.
Typ
e Field Test Site Material Regression Equation
Coefficient of Multiple
Determination (R2) for composite modulus as
a function of percent compaction and water content
relative to optimum
I-84 Crushed Gravel Mc = -294.49+4.728(PC)+8.833(RWC) 0.679
Sand Mc = -87.276+2.828(PC)+6.331(RWC) 0.602 Route 25 Gravel Mc = 652.69-4.249(PC)+1.504(RWC) 0.001 Bas
e
Combined --- Mc = -411.26+5.454(PC)–2.757(RWC) 0.823 I-84 Subgrade* Mc = -1167.8+8.798(PC)+27.086(WC) 0.497 Sub-
grade CPR Subgrade Mc = 174.53+1.038(PC)–6.947(RWC) 0.044 * - based on dry density and water content. Mc = composite modulus; PC = percent compaction; RWC = relative water content.; WC = water content Note: Range of water contents for the base course results are 3 to 9.5% dry of optimum water content so the regression equations should be used with caution for higher water contents.
228
variables. This was only slightly higher than when considering percent compaction
alone, but as noted previously the field sites were all dry of optimum and the range of
water contents was small. The influence of water content may have been higher is a
greater range of water contents was tested at the field sites.
The relationship between the combined regression equations based on the
laboratory and field results was explored by plotting the predicted composite modulus
versus percent compaction for water contents of 4% dry of OWC, at OWC, and 4% wet
of OWC as shown in Figures 5.41 through 5.43. It is seen that there is reasonable
agreement between the composite moduli predicted by the three regression equations.
Moreover the slopes for the lab (select) and field equations are very similar.
The relative importance of percent compaction and water content on the predicted
composite modulus was examined using the regression equation based the combined lab
results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel. This
equation was selected since it has a higher correlation coefficient than the equation based
on all the lab results and it is based on data with a much larger range of water contents
that at the field sites. The comparison is shown in graphical form in Figure 5.44. It is
seen that a change in percent compaction from 90 to 100% would increase the composite
modulus by about 50 MPa whereas increasing the water content from 4% dry of optimum
to 4% wet of optimum would decrease the composite modulus by about 60 MPa. Thus,
percent compaction and water content relative to optimum have a similar significance on
the predicted composite modulus.
229
85 90 95 100 105Percent Compaction (AASHTO T-180)
0
100
200
300
Com
posi
te M
odul
us (M
Pa)
Lab (all) OWC-4%Lab (select) OWC-4%Field OWC-4%
Lab (select) includes only results for CTcrushed gravel, NH Sand, & Wardwell Gravel
Figure 5.41 Composite modulus predicted by regression equations at 4% dry of
optimum.
85 90 95 100 105Percent Compaction (AASHTO T-180)
0
100
200
300
Com
posi
te M
odul
us (M
Pa)
Lab (all) OWCLab (select) OWCField (OWC)Extrapolated data
Lab (select) includes only results for CTcrushed gravel, NH Sand, & Wardwell Gravel
Figure 5.42 Composite modulus predicted by regression equations at optimum.
230
85 90 95 100 105Percent Compaction (AASHTO T-180)
0
100
200
300
Com
posi
te M
odul
us (M
Pa)
Lab (all) OWC+4%Lab (select) OWC+4%Field OWC+4%Extrapolated data
Lab (select) includes only results for CTcrushed gravel, NH Sand, & Wardwell Gravel
Figure 5.43 Composite modulus predicted by regression equations at 4% wet of
optimum.
85 90 95 100 105Percent Compaction (AASHTO T-180)
0
100
200
300
Com
posi
te M
odul
us (M
Pa)
Lab (select) OWC-4%Lab (select) OWC-2%Lab (select) OWCLab (select) OWC+2%Lab (select) OWC+4%
Lab (select) based on only results for CTcrushed gravel, NH Sand, & Wardwell Gravel
Figure 5.44 Effect of percent compaction and water content on predicted composite
modulus based on laboratory results for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel.
231
5.3.4 Additional Factors Influencing Composite Modulus
The stiffness characteristics of compacted cohesionless materials are not only
influenced by percent compaction and water content, but also by mineralogical
composition, size and gradation of the individual particles, and shape of the individual
particles (Langfelder and Nivargikar, 1967). Aggregate characteristics for laboratory
samples are provided in Table 5.10.
Laboratory results were separated by sample, degree of compaction, and water
content relative to optimum. Trials with relative water contents outside the range of ±
3% of optimum were removed. The remaining results were separated into three groups
based on percent compaction. The low degree of compaction refers to those trials which
the percent compaction was less than or equal to 92%. Moderate degree of compaction
refers to those trials for which the percent compaction was between 93% and 97%.
Finally, high degree of compaction refers to those trials for which the percent compaction
was greater than or equal to 98%. Average composite moduli for these categories, as
determined from the Prima 100 PFWD, are presented in Table 5.10.
Table 5.10 Comparison of average composite moduli associated with varying degrees of compaction.
Prima 100 PFWD Composite Modulus (MPa)
Sample
Coefficient of
Uniformity (Cu)
PercentFines (%)
Low Degree of Compaction
(E ≤ 92%γdmax)
Moderate Degree of Compaction
(93% ≤ E ≤ 97%γdmax)
High Degree of Compaction
(E >= 98%γdmax)Connecticut
crushed gravel 18 5.2 114 162 145 New Hampshire
sand 5 2.8 75 137 NA New Hampshire
gravel 20 2.9 126 94 91 OJF gravel 10 4.5 NA 61 71
Wardwell gravel 47 6.0 105 108 135 NA – percent compaction was not achieved for given range.
232
In general terms, for the low degree of compaction, larger composite moduli are
associated with higher coefficient of uniformity (Cu). This trend is also evident for
moderate and high degrees of compaction with the exception of New Hampshire sand
which produced the second largest composite moduli. High values of Cu indicate that a
material is well graded. Fine grained particles present in the well graded materials fill the
voids between larger particles allowing for a more compact soil fabric and a higher
modulus.
The shape of the soil particles may have some effect on the composite modulus.
Samples with subangular and angular particles appeared to produce somewhat larger
composite moduli than those samples with rounded and subrounded particles. A visual
description of particle shapes is provided in Figure 5.45. The most distinct difference is
seen when comparing New Hampshire sand (Figure 5.46) and Connecticut crushed gravel
(Figure 5.47). The Connecticut crushed gravel sample tended to have higher composite
moduli as shown in Table 5.10. The combination of subangular and angular shaped
particles and a higher Cu appears to produce higher composite moduli.
5.4 COMPARISON OF PORTABLE DEVICES
The repeatability of results from two different Prima 100 PFWDs as well as a
comparison between results from the Prima 100 PFWD and Clegg Impact Hammer (CIH)
are examined in this section. A description of the portable devices is provided in Section
2.2.1. A comparison of results from two Prima 100 PFWDs is presented first followed by
a comparison between the two different PFWD devices. The effect of operator technique
is discussed last.
233
Figure 5.45 Typical shapes of coarse grained bulky particles (Holtz and Kovacs, 1981).
Figure 5.46 Poorly graded New Hampshire sand.
234
Figure 5.47 Well graded Connecticut crushed gravel.
5.4.1 Prima 100 PFWD Comparison
The results from two Prima 100 PFWDs were compared on gravel samples
obtained from Owen J. Folsom & Sons, and Robert Wardwell & Sons Inc. One Prima
100 PFWD was purchased by the UMaine in the spring of 2003 and the other was
purchased by the United States Forest Service (USFS) in late fall 2003. During
laboratory tests, USFS PFWD measurements were performed first, followed by the
UMaine PFWD. Both Prima 100s are shown in Figure 5.48. At each test point, 18 drops
(six drops at three different heights) were completed with each device. Additional testing
details can be found in Section 3.5.2.3. Composite moduli derived from both devices are
plotted versus drop number. This is shown in Figures 5.49 and 5.50.
235
USFS Prima 100 composite moduli are less than UMaine Prima 100 composite
moduli for each drop height. In general, the first drop at each drop height was less than
the remaining drops for the same height. This is true for both devices. Prior to the first
drop, nonuniform contact exists between the loading plate and the gravel surface. Once
the first drop occurs, the surface is compacted and more uniform contact between the
surfaces develops. The test point becomes more compact with successive drops.
UMaine Prima 100 composite moduli are greater, at least in part, due to the amount of
additional compaction the test point has undergone due to testing with the prior testing
with the USFS Prima 100.
Figure 5.48 Testing with USFS and UMaine Prima 100 PFWDs.
A plot of the composite modulus determined from the USFS PFWD versus the modulus
from the UMaine PFWD at the 850 mm (33.5 in.) drop height is shown in Figure 5.51.
236
The plot confirms that the UMaine Prima 100 produced slightly higher moduli than the
USFS Prima 100. A regression analysis yielded a correlation coefficient of 0.954. This
shows that two different Prima 100 units give results that correlate very well with each
other, even though the modulus values are slightly different. To determine the percent
difference between the two units, the regression equation shown in Figure 5.51 was used
to compute the USFS moduli corresponding to three different UMaine moduli. The
results are shown in Table 5.11. This shows that the UMaine modulus is 18 to 20 percent
larger than the USFS modulus. As noted previously, this could be due in part to the
higher compaction for the tests with the UMaine device. Alternately, the UMaine device
could be biased to give results that are slightly higher than the USFS device.
6 12 18 24 30 3Drop Number
60
40
80
120
160
Com
posit
e M
odul
us (M
Pa)
USFS Prima 100 PFWD UMaine Prima 100 PFWD
850
mm
dro
p he
ight
850
mm
dro
p he
ight
630
mm
dro
p he
ight
630
mm
dro
p he
ight
420
mm
dro
p he
ight
420
mm
dro
p he
ight
Figure 5.49 Change in composite modulus with subsequent drops for two PFWDs at the same test point for OJF gravel at 100% compaction and optimum water content (TP #1).
237
6 12 18 24 30Drop Number
360
40
80
120
160
Com
posit
e M
odul
us (M
Pa)
USFS Prima 100 PFWD UMaine Prima 100 PFWD
850
mm
dro
p he
ight
630
mm
dro
p he
ight
420
mm
dro
p he
ight
850
mm
dro
p he
ight
630
mm
dro
p he
ight
420
mm
dro
p he
ight
Figure 5.50 Change in composite modulus with subsequent drops for two PFWDs at the same test point for Wardwell gravel at 100% compaction and optimum water content (TP #1).
0 40 80 120 160 200UMaine Prima 100 PFWD Composite Modulus (MPa)
0
40
80
120
160
200
USF
S Pr
ima
100
PFW
D C
ompo
site
Mod
ulus
(MPa
) Equation Y = 0.844*X - 3.604Coef of determination, R-squared = 0.954
Figure 5.51 Comparison of USFS and UMaine Prima 100 PFWD composite moduli.
238
Table 5.11 Comparison of composite moduli from USFS and UMaine Prima 100 PFWDs.
USFS Prima 100 PFWD
Composite Modulus (MPa)
UMaine Prima 100 PFWDComposite Modulus
(MPa)
Percent Difference
(%) 64 80 20 98 120 18 131 160 18
5.4.2 Clegg Impact Hammer
The Clegg Impact Hammer (CIH) was used on all laboratory samples. CIH
measurements were taken last, after both the USFS and UMaine PFWDs. Six
measurements were taken at each location. Additional information pertaining to testing
techniques is provided in Section 3.5.2.3. Composite moduli derived from each device
are plotted against drop number. Results are shown in Figure 5.52.
Prima 100 composite moduli are greater than composite moduli from CIH
measurements. In general, the first drop with the CIH produced moduli that were less
than those from subsequent measurements. Moreover, the CIH moduli tended to increase
with each subsequent drop. It appeared that a shallow bearing capacity failure occurred
upon impact of the falling mass, displacing material down and to the side of the contact
surface. This is shown in Figure 5.53. This could be an explanation of why the CIH
modulus increased with each subsequent drop. Moreover, the additional vertical
displacement induced by bearing capacity failure could contribute to a lower CIH moduli.
A plot of Prima 100 PFWD moduli versus CIH moduli is shown in Figure 5.54. A
regression analysis yielded a correlation coefficient of 0.483, significantly greater than
that obtained during spring thaw monitoring during the spring of 2003 at the USFS
Parking Lot. The coefficient of simple determination (r2) was 0.230.
239
Table 5.12 Comparison of Prima 100 PFWD composite moduli and CIH moduli.
Prima 100 PFWD Composite Modulus
(MPa)
Clegg Impact HammerModulus
(MPa)
Percent Difference
(%) 75 8 89 150 26 83 225 43 81
6 12 18 24 30 36 4Drop Number
20
40
80
120
160
Com
posit
e M
odul
us (M
Pa)
USFS Prima 100 PFWD UMaine Prima 100 PFWD
850
mm
dro
p he
ight
630
mm
dro
p he
ight
420
mm
dro
p he
ight
850
mm
dro
p he
ight
630
mm
dro
p he
ight
420
mm
dro
p he
ight
CIH
Figure 5.52 Change in composite modulus with subsequent drops for different devices
at the same test point for OJF gravel at 100% compaction and optimum water content (TP #1).
240
Figure 5.53 Clegg Impact Hammer measurement on New Hampshire sand.
0 50 100 150 200 250Prima 100 PFWD Composite Modulus (MPa)
0
50
100
150
200
250
Cle
gg Im
pact
Ham
mer
Mod
ulus
(MPa
)
Equation Y = 0.233*X - 9.139Coef of determination, R-squared = 0.483
Figure 5.54 Comparison of Clegg Impact Hammer and UMaine Prima 100 PFWD
composite moduli.
241
5.4.3 Effect of Operator Technique
The effect of operator technique was investigated by comparing the results
obtained by five different operators, two of which were Maine Department of
Transportation technicians. The remaining three operators were made up of
undergraduate and graduate research assistants. One of the operators is shown in Figure
5.55. Each user trial was completed on Wardwell gravel sample which was several
percentages above optimum water content. Initial instructions were given by the
principal investigators to each of the operators prior to use. Each operator completed six
measurements at each of the five test locations. The results are compared to determine to
what extent individuals operators could produce similar results. This is shown in Table
5.13. Average composite moduli for each user decreased as the testing progressed. In
other words, average composite moduli for operator #2 was less than operator #1,
operator #3 less than operator #2, and so on. One possible explanation for this is that
since the trials were performed on a sample wet of optimum, successive measurements by
multiple operators locally increased pore water pressure at each of the test locations. As
pore water pressure is increased, modulus (stiffness) is reduced. Minimal elapsed time
between users did not allow for dissipation of pore water pressure. Based on visual
observation of the operators, none appeared to have difficulties in using the device. Their
technique differed only slightly from that used by the primary investigators. As a result,
it appears that only minimal training would be required in order to produce acceptable
results.
242
Figure 5.55 MaineDOT representative (operator #2) performing laboratory PFWD measurements.
Table 5.13 Comparison of Prima 100 PFWD composite moduli determined by different users.
Test Point
Operator #1
Operator #2
Operator#3
Operator#4
Operator#5
Average & (Standard Deviation)
1 30 23 26 31 25 27.0 (3.3) 2 34 33 33 35 25 34.6 (1.6) 3 29 31 22 12 13 21.4 (8.9) 4 22 28 13 14 12 17.6 (6.9) 5 29 27 27 22 19 24.8 (4.0)
Average 28.8 28.4 24.2 22.8 18.8 25.1 Std. Dev. 4.3 3.9 7.4 10.1 6.3 4.9
5.5 RECOMMENDATIONS
Despite the importance of modulus, some aspects of pavement construction and
management are still based on measurement of other parameters that are not directly
243
connected with long-term performance or, even less desirable, on empirical based
judgments. One critical area that does not currently make use of modulus is evaluating
the adequacy of subgrade and base compaction during construction.
Based on the results of this research the tentative procedure given below is
recommended for using the Prima 100 PFWD to monitor compaction of granular base
courses. The procedure is based on the observation that there is a rough equivalency
between percent compaction and composite modulus for granular base at optimum water
content. Correction factors are recommended to correct the composite modulus measured
at the field water content to the equivalent value at optimum water content. The
regression equation for the combined results for Connecticut crushed gravel, New
Hampshire sand, and Wardwell gravel (Table 5.8) was used to derive the
recommendations. This equation was used since it had a higher correlation coefficient
than the regression that included all five laboratory samples combined and it had a larger
range of water contents than the field samples.
The target composite modulus at optimum water content should be chosen based
on Table 5.14 that gives a rough equivalency with percent compaction based on
AASTHO T-180. Composite moduli measured in the field should be corrected to the
equivalent composite modulus at optimum water content by adding the factors given in
Table 5.15. Thus, it is necessary to determine the field water content relative to OWC to
apply this procedure. Possibilities for measuring the water content include oven drying,
pan drying, Speedy Moisture Meter©, time domain reflectometry, or nuclear density
meter in backscatter mode. The researchers caution that the values given in Tables 5.14
244
and 5.15 are based on a limited dataset. It is recommended that these equivalences be
confirmed for additional materials used by individual state DOTs.
Table 5.14 Tentative equivalences between percent compaction and composite modulus at optimum water content for base and subbase course aggregate.
Percent
Compaction based onAASTHO T-180 (%)
Equivalent Prima 100 PFWDComposite Modulus (MPa) at
Optimum Water Content 90 92 95 115 98 130 100 139
Table 5.15 Factor to correct composite modulus measured at field water content to equivalent value at optimum water content.
Water Content Relative to Optimum
Correction Factor to be Added to Composite
Modulus (MPa) Measured at Field Water Content
-4% -31 -3% -23 -2% -15 D
ry o
f O
WC
-1% -8 At OWC 0
+1% 8 +2% 15 +3% 23 W
et o
f O
WC
+4% 31
Subgrade soils were tested in the field as part of this project. However, the range
of water contents was small. Thus, there is insufficient data to draw conclusions on
equivalent composite moduli and the very important effect of water content at the time of
testing for subgrade soils.
245
Field testing techniques for monitoring compaction of aggregates using the Prima
100 PFWD have been developed. These recommendations are based on the experiences
of the researchers in using the Prima 100 PFWD as discussed in previous sections. The
procedure is similar to that used for testing for thaw weakening discussed in Chapter 4.
The Prima 100 PFWD should be setup with the 850 (33.5 in.) drop height, 20 kg (44 lb)
drop weight, and 300 mm (12 in.) diameter loading. The Personal Digital Assistant
(PDA) based recording software should be setup using the input parameters presented in
Table 5.16. Six measurements should be taken at each test location. The first
measurement should be discarded and the average of the remaining five should be used as
the modulus at that test location. In addition, three locations should be tested within a 3
m (10 ft) diameter area, the average of which will provide a representative value for that
particular station.
Table 5.16 Prima 100 PFWD input parameters.
Setup MenuItem
Input Parameter Value
Pretrig time (ms) 10* Pulsebase (%) 24* Trigger
Trig Level (kN) 0.90* View Sample Time (ms) 60*
Load Plate Radius (mm) 150 Number of sensors 1 Mechanical
D(1) offset (cm) 0 Poisson’s Ratio 0.35** Formula
Stress Distribution 2.67 * - default values. ** - Huang (2004)
246
Further research is recommended to validate and refine this procedure for a wider
range of soil types. This should be based on measurements obtained at field test sites.
Manual water content and sand cone tests should be performed to check the
measurements obtained with the NDM. Samples should be obtained for gradation,
maximum dry density, and optimum water content determination.
5.6 SUMMARY
This chapter presents the analysis and results of the field and laboratory
evaluation of the PFWD as an alternative to traditional compaction control devices. As
part of this effort, the relationship between PFWD composite moduli, percent
compaction, and water content relative to optimum for soil types representative of New
England base and subbase aggregates was explored. PFWD and NDM measurements
were taken at five field test sites during the summer and fall of 2003. In addition,
laboratory tests were completed on five different samples during the summer of 2004.
For the laboratory tests, each soil sample was compacted in lifts in the 1.8 m x 1.8
m x 0.9 m (6 ft x 6 ft x 3 ft) deep test box initially at a low density. Measurements were
taken, then the sample was removed, reconditioned to the desired water content, and
recompacted to a higher density and the measurements repeated. Tests were performed
percent compactions ranging from 86 to 103 percent and water contents ranging from 3.5
percent dry to 10.2 percent wet of optimum. In total, 29 combinations of water content
and density were tested in the laboratory.
The relationship between composite modulus and percent compaction was
explored. For four out of the five materials tested, there was a general trend of increasing
247
composite modulus with increasing percent compaction. New Hampshire gravel
exhibited the opposite trend. With the exception of the New Hampshire Sand, the
regression coefficients were less then 0.5 indicating poor correlation. Combining all the
results yielded a correlation coefficient of 0.045, indicating no correlation. However,
including only the results for Connecticut crushed gravel, New Hampshire sand, and
Wardwell gravel resulted in a correlation coefficient of 0.35, which still indicates a poor
correlation. Regression coefficients were slightly higher when separate correlations were
developed for samples dry and wet of optimum, however the r2 were still less than 0.5.
Results from the field test sites also indicate that as the degree of compaction increases,
composite modulus increases. In general, correlation coefficients were greater for field
test results compared to laboratory test results. Combining the results for the three base
materials tested in the field, resulted in a correlation coefficient of 0.818, which is a
relatively strong correlation. However, the significance of this correlation is diminished
by the fact the water content at all the field sites was dry of optimum. The Prima 100
PFWD also proved adequate in measuring time dependent increases in composite
modulus for asphalt stabilized base products tested on Route 201 in The Forks, Maine,
and at Commercial Recycling Systems in Scarborough, Maine.
Laboratory results also show that there is a general trend that the composite
moduli tends to decrease as water content increases. Correlation coefficients ranged from
0.003 (Connecticut Crushed Gravel) to 0.814 (Wardwell Gravel). The low correlation
coefficients for several of the samples are due in part to the role that percent compaction
plays in the composite modulus, which is not accounted for when only water content is
considered. Combining at the laboratory results yielded a correlation coefficient of 0.285
248
which indicates poor correlation. For measurements taken at field sites the correlation
coefficient ranged from 0.008 (Route 25 Gravel) to 0.521 (Route 25 Sand). However,
water contents measured at field sites were generally drier than -3% of the OWC and in
some instances were as low as -9% of the OWC, which are significantly different from
water contents obtained during laboratory tests.
It is known that stiffness characteristics of compacted cohesionless materials are
not only influenced by percent compaction and water content, but also by mineralogical
composition, size and gradation of the individual particles, and shape of the individual
particles (Langfelder and Nivargikar, 1967). The effect of these factors on composite
moduli was difficult to quantify and future research in this area is recommended.
Multivariable linear regression analyses (Neter, et al., 1982) were used to
determine the best fit line for composite modulus as a function of percent compaction and
water content. The R2 for the laboratory materials ranged from 0.141 (Connecticut
crushed gravel) to 0.867 (Wardwell gravel). Combining all laboratory samples produced
an R2 of 0.326. However, including only laboratory results for Connecticut crushed
gravel, New Hampshire sand, and Wardwell gravel increased the R2 to 0.624. This
indicates that 62% of the variation in composite modulus is explained by the percent
compaction and water content relative to optimum. The R2 for the field materials ranged
from 0.001 (Route 25 gravel) to 0.679 (I-84 crushed gravel). Combining the three field
sites where granular base was tested yielded an R2 of 0.823, which indicates a reasonably
strong correlation of composite modulus with percent compaction and water content,
independent of the type of material tested. However, the water contents for the field sites
were all dry of optimum which may limit the significance of this result. The multi-
249
variable linear regressions based on the three laboratory samples indicated above and
field results yielded predicted composite modulus at 95% percent compaction that were
agreed within 20% which is reasonable agreement.
The results from two Prima 100 PFWDs were compared on gravel samples
obtained from Owen J. Folsom & Sons and Robert Wardwell & Sons Inc. The UMaine
Prima 100 produced slightly higher moduli than the USFS Prima 100. The USFS PFWD
was used first, completing six measurements at three different drop heights, at each of the
five test locations before the process was repeated with the UMaine PFWD. The
differences in moduli could, in part, be due to additional compaction test points
underwent as a result of using the USFS PFWD first. A regression analysis yielded a
correlation coefficient of 0.954. This shows that two different Prima 100 units give
results that correlate well with each other, even though the modulus values are slightly
different.
The Prima 100 PFWD and Clegg Impact Hammer (CIH) were used on all
laboratory samples. Prima 100 composite moduli were greater than composite moduli
derived from CIH measurements. The occurrence of a shallow bearing capacity failure
caused by the impact of the CIH could help to explain the large differences between
moduli from the devices. Additionally, the first drop with the CIH produced moduli that
were less than those derived from subsequent measurements. Finally, the CIH moduli
tended to increase with each subsequent drop.
The effect of operator technique was investigated by comparing the results
obtained by five different operators on one trial of Wardwell gravel (90% compaction,
+10% wopt). Average composite moduli for each user decreased as testing progressed
250
which could partially be caused by the presence of excess water in the material at the
time of testing, as minimal time was allotted for dissipation of pore water pressure
between tests.
Recommendations were made for utilizing the Prima 100 PFWD as a tool to
monitor compaction. Tentative equivalences between percent compaction and composite
modulus for base course aggregates at optimum water content were provided. The
recommendations were based on a multi-variable linear regression on laboratory results
for Connecticut crushed gravel, New Hampshire sand, and Wardwell gravel. In addition,
correction factors to correct composite moduli measured at the field water content to the
equivalent value at optimum were proposed. Additional study on a wider range of base
course aggregates and water contents is recommended.
251
CHAPTER 6
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
6.1 SUMMARY
Vehicles can cause significant damage to roads that are weakened during the spring
thaw. To minimize damage, many road maintenance agencies impose load restrictions on
selected roads during damage susceptible periods. Although, the maximum allowable load
and the duration of the reduced load period vary widely among agencies, they try to strike a
balance between minimizing the disruption to the local economy caused by the load
restrictions and minimizing road damage.
The decrease in the stiffness of the pavement system during the spring thaw is a key
factor leading to pavement damage. The term stiffness is often used interchangeably with
elastic modulus, resilient modulus, or in some cases, simply modulus. Stiffness can be
monitored during spring thaw and through recovery using a Falling Weight Deflectometer
(FWD). However, FWD purchase, operation, and maintenance is expensive. Moreover,
even if a state owns a FWD, it can only cover so many roads within a given time frame. As a
result, determining when the road has thawed to the point where a load restriction is needed
and when the road has recovered sufficient strength to remove the restriction is often left to
personal experience and subjective judgment.
Irrespective of the seasonal changes in modulus, the overall stiffness of compacted
base and subgrade layers has a significant effect on pavement life. The current practice
during construction is to achieve a high modulus by reaching a specified percent compaction.
However, current techniques for monitoring compaction exhibit numerous shortcomings.
252
Prior to the 1980’s, the sandcone was widely used to measure density and provide quality
control for construction. However, the sandcone test is labor intensive and time consuming.
In practice, it has been replaced by the nuclear moisture density gauge, or densimeter.
Although considerably faster than the sandcone, the densimeter’s radioactive source
constrains use and handling. Because of the problems, compaction conformance is
sometimes left to subjective judgment. This may compromise the integrity of the finished
project resulting in reduced service life and increased maintenance costs. Moreover, none of
the methods measure stiffness directly.
The objective of the study were to address the concerns discussed above. Specific
objectives are listed below.
1. Evaluate the portable falling weight deflectometer (PFWD) as a means of
optimizing timing for load restriction placement and removal on secondary
roads in New England.
2. Develop guidelines for PFWD use on pavement structures typical of New
England low volume roads.
3. Evaluate the effectiveness of the PFWD as a means for monitoring
compaction, density, or bearing capacity at construction sites. This includes
developing correlations between PFWD results and percent compaction for a
range of soils.
4. Develop guidelines, including acceptance and testing protocols, determined
via testing and subsequent conventional and statistical analyses.
253
5. Compare the results from different PFWD’s and several alternative devices
for measuring the degree of compaction of highway subgrade soils and
base/subbase coarse aggregates.
6.1.1 Literature Review
Portable falling weight deflectometers (PFWD) have been developed by several
manufacturers to measure the in situ stiffness of construction layers including subgrades,
base courses, and pavements. For paved roads, moduli or deflection results determined by
PFWDs were compared by several investigators to values determined by FWDs and
Benkelman Beams. The Loadman PFWD was used for most of these studies. In general, the
comparisons showed marginal correlation coefficients (r2) less than 0.5, however one study
obtained an r2 of 0.86 for a correlation between moduli determined by the FWD and PFWD.
The PFWD generally produced higher modulus values than the FWD, possibly due to the
smaller depth of influence. Several investigators reported that the zone of influence for the
PFWD lies primarily between one and two loading plate diameters. Large aggregate
particles beneath the loading plate of the portable devices also affect the results, as the
particles increase the resulting modulus values. Some investigators imply that the PFWD is
better suited to roads with thin pavements.
For unbound layers, PFWD results were compared to results from FWD, plate
bearing test, Clegg Hammer, surface stiffness gage, and Benkelman Beam. PFWD from
several manufacturers were tested including Handy, Loadman, Prima 100, and German
Dynamic Plate Tester. In general, each device could detect changes in soil stiffness. Several
studies found good correlation between moduli from PFWD and FWD with r2 between 0.31
and 0.99 with most values greater than 0.5. The Clegg Hammer generally correlated poorly
254
with results from other devices. reasonably with other devices when testing on unbound
layers. A limited number of studies examined correlations between PFWD moduli and
percent compaction. Correlations were generally poor.
The PFWD has the potential to track seasonal stiffness variations in paved and
unpaved low volume roads. Pavement modulus is a key parameter in determining damage-
susceptibility of pavements. Pavements in areas with seasonal freezing and thawing often
undergo frost heave and thaw weakening in addition to load-induced pavement distress. To
minimize damage, many road maintenance agencies impose load restrictions during damage-
susceptible periods. This can be monitored during spring thaw and through recovery using a
traditional FWD to assist in determining when to place and remove the restrictions.
However, the initial investment in purchasing a FWD, as well as high operation and
maintenance costs limits its use. As a result, determining when the road has thawed and
recovered sufficient strength to remove the restriction is left to personal experience and
subjective judgment.
Limited studies have been conducted to evaluate the PFWD as a tool for tracking
seasonal stiffness variations. Correlation coefficients relating the PFWD and the Benkelman
Beam were generally high. For thin pavement sections, the PFWD did adequately follow the
seasonal stiffness variations.
An increasing number of studies have been conducted to develop quantitative
techniques that may be used to better determine when load restrictions should be placed and
removed. A majority of state and local agencies use subjective techniques, such as
observation, to both place and remove load restrictions. Fewer departments were using
quantitative methods to place load restrictions, however, an even smaller number use the
255
same methods to remove load restrictions. Additionally, some simply keep restrictions in
place for a specific length of time. Studies conducted in Washington and Minnesota have
attempted to use air temperature to determine when to place and remove the restrictions, and
appear to have worked adequately for the conditions in their respective states. Additional
work done in Washington has suggested that using deflection data to aid in load restriction
placement and removal can be done and recommends that during the spring thaw, once the
deflections reach 40 to 50% of their fully recovered values, weight restrictions should be
placed and removed.
6.1.2 Field & Laboratory Test Protocol
The performance of seven paved and three gravel surfaced roads were monitored
during the spring of 2004. Test sites were located in Maine, New Hampshire, and Vermont.
Two additional sites in Northern Maine were used for testing on one day as part of an
ongoing MaineDOT research project. The test sites varied in asphalt thickness, subbase
thickness and type, and subgrade type.
Instrumentation was used to quantify the condition of the test sections on days when
measurements were made. Thermocouples, thermistors, and frost tubes were used at selected
sites to monitor the advance and retreat of freezing conditions during late winter and spring
months. Vibrating wire and standpipe piezometers were installed at selected sites to monitor
pore water pressures in the subbase, and at some sites, subgrade layers. Time Domain
Reflectometry (TDR) probes were used to monitor water content through the spring thaw and
recovery periods at some sites. Instrumentation was used to examine the extent to which the
road had thawed and provided the context for interpretation of PFWD and FWD results.
256
Manual instrumentation readings were taken approximately weekly through the spring thaw
and recovery periods. At selected sites, instrumentation was read automatically by data
acquisition systems and was downloaded approximately weekly.
Prima 100 PFWD and traditional FWD measurements were taken at a minimum of
eight locations at each test site. Measurements were taken approximately weekly during the
spring thaw period. In addition, Loadman PFWD measurements were taken at spring thaw
test sites in Rumney, New Hampshire. Clegg Impact Hammer and Humboldt Soil Stiffness
Gauge measurements were taken at the United States Forest Service (USFS) Parking Lot
during the spring of 2003 and 2004. With the Prima 100 PFWD, six measurements were
taken at each of three different drop heights, at each test location. The drop heights were
approximately 850, 630, and 420 mm (34, 25, and 17 in.). Deflection sensors were used with
spacing as follows (as measured from the center of the loading plate): 0, 207, and 407 mm (0,
8, and 16 in.). The PFWD measurements were taken utilizing a 20 kg (44 lb) drop weight
and a 300 mm (11.8 in.) loading plate. In all cases, the first reading was neglected and the
average of the remaining five was used for analysis and comparison. In addition, five
Loadman PFWD, four Clegg Impact Hammer, and one Soil Stiffness Gauge measurement
was taken at each test location. The MaineDOT provided a FWD for seasonally posted roads
in Maine. The United States Army Corps of Engineers Cold Regions Research and
Engineering Laboratory (CRREL) provided a FWD for test sites in Rumney, New
Hampshire. The Vermont Agency of Transportation (VAOT) provided a FWD for
seasonally posted low volume roads in Vermont. MaineDOT provided backcalculation of
FWD data from field test sites in Maine using DARWin. The researchers completed
257
backcalculation procedures for FWD data obtained from New Hampshire and Vermont field
test sites using Evercalc.
Five field test sites in Maine, New Hampshire, and Connecticut were used to evaluate
the effectiveness of the PFWD as a tool to monitor compaction. Different material types
were tested at each field site. Tests were performed at a minimum of 12 locations, utilizing
both the Prima 100 PFWD and Nuclear Moisture Density Gauge (NDM) (AASHTO T 238).
With the PFWD, six measurements were taken at each test location utilizing the 20 kg drop
weight, 300 mm diameter loading plate, and 850 mm drop height. Only the deflection sensor
integral to the loading plate was used. The Connecticut Department of Transportation and
the New Hampshire Department of Transportation provided a NDM for field test sites in
their respective states. NDM measurements were taken at depths of 203, 152, 102, 51, and 0
mm (8, 6, 4, 2, and 0 in.) at each test location. Samples were taken at each site for sieve
analysis, maximum dry density, and optimum water content determination. Tests were
performed in accordance with AASHTO test procedures.
The primary purpose of the laboratory component of this project was to determine a
relationship between PFWD results and percent compaction under controlled conditions.
The large-scale laboratory study to correlate PFWD results to percent compaction was
constructed in the geotechnical research laboratory at the University of Maine. The tests
were conducted in a 1.8 m x 1.8 m x 0.9 m (6 ft x 6 ft x 3 ft) deep test container. The bottom
203 mm (8 in.) of material met MaineDOT Type D aggregate specifications. This material
was kept in place throughout all laboratory tests. Five different material types were used to
fill the remaining height of the test box. Material was added to the container in
approximately 152 mm (6 in.) lifts. Each lift was compacted using a hand tamper and
258
electric jackhammer with a modified flat plate attachment. Each aggregate was compacted in
the container to approximately 90, 95, and 100% of the maximum dry density (AASHTO T
180). The effect of water content was determined at 95% of the maximum dry density.
Measurements were taken at optimum water content as well as ± 3% of the optimum water
content. Once all the material was compacted in the test container, several portable testing
devices were used. Prima 100 PFWD, Clegg Impact Hammer, NDM, and Dynamic Cone
Penetrometer (DCP) tests were performed at multiple locations. Prima 100 PFWD and Clegg
Impact Hammer measurements were taken in the same manner as was done for the spring
thaw portion of the research. In addition, one sand cone test was completed, and two water
content samples were taken for each trial for comparison to NDM measurements.
6.1.3 Spring Thaw Monitoring
6.1.3.1 Instrumentation Measurements
Subsurface temperatures were measured at each field site during the end of the
freezing season, throughout the thawing period, and into the recovery period. Measurements
taken at asphalt surfaced test sites indicated freezing temperatures penetrated to their
maximum depths between February 17 and March 24, 2004. Maximum depths ranged from
a low of 866 mm (34 in.) at Stinson Lake Road to a high of 1930 mm (76 in.) at Route 1A
(Section D-2). Complete thaw occurred at all test sites between mid-March and mid-April.
Measurements taken at gravel surfaced test sites indicated freezing temperatures penetrated
to their maximum depths between March 1 and April 21, 2004. Maximum depths ranged
from a minimum of 1128 mm (44 in.) at the USFS Parking Lot to a maximum of 2134 mm
259
(84 in.) at Crosstown Road. Complete thaw had occurred at all sites between early April and
mid May.
Pore water pressures were measured in the subbase layer at each field site during the
thawing and into the recovery period. At some sites, pore water pressures were also
measured in the subgrade layer. The highest water levels observed in standpipe piezometers
roughly correspond to the date of complete thaw for that particular site. This is true for both
asphalt and gravel surfaced test sections. TDR moisture sensors used at the USFS Parking
Lot and Stinson Lake Road also indicate an increase in moisture content at or near the date of
complete thaw. The relationship between manual vibrating wire piezometer measurements
and thawing at the Route 126 and Route 1A sites is unclear.
At the Route 126, Sections 3 and 8, vibrating wire piezometer readings were
monitored hourly by a datalogger. At the initiation of thawing, subgrade porewater pressures
in these sections were negative. Once thawing commented, the subgrade porewater pressure
increased reaching a maximum head of 1.2 m (4 ft) about a month after complete thawing.
In the subbase at Section 3, the head was near zero from the initiation of thawing through the
end of monitoring in mid-June. In Section 8, the subbase pore water pressure was about -0.6
m (-2 ft) from the initiation of thawing through the end of monitoring.
Overall, the piezometer results indicate that at most sites higher porewater pressures
in the subgrade and subbase soils were associated with the thawing period. This is a factor
that could contribute to reduction of pavement stiffness during the spring thaw.
260
6.1.3.2 Seasonal Stiffness Variations
For each test site, Prima 100 PFWD composite modulus, and where it is available,
FWD asphalt, subbase, subgrade, and composite modulus and Loadman PFWD composite
modulus values were plotted versus date. In general, for asphalt surfaced test sites, the
moduli are high when the pavement section is frozen and during the early part of the period
when section is partially thawed. At some field sites there are significant differences in
moduli from nearby test locations and from one week to the next. This variability is more
apparent in gravel surfaced test sites compared to asphalt surfaced test sites. For both asphalt
and gravel surfaced test sites, composite moduli generally decreased as thawing progressed.
It was anticipated that a distinct minimum would occur before increasing through the
recovery period. However, this was only evident at the Buffalo Road, USFS Parking Lot,
Knapp Airport Parking Lot, Crosstown Road, and to a lesser extent Stinson Lake Road. At
the remaining sites, the composite modulus that was reached during the spring thaw was
about the same as, or in some, cases greater than the values measured during the summer.
FWD derived layer moduli confirm these observations. In general, portable and traditional
FWD moduli follow similar trends for both asphalt and gravel surfaced test sites through the
monitoring period. Thus, the PFWD and FWD would be equally as effective in monitoring
stiffness change during the spring thaw.
6.1.3.3 Comparison of Portable and Traditional FWD Moduli
The degree of correlation between composite moduli backcalculated using FWD and
Prima 100 PFWD results were investigated. This was done for five sites in Maine where the
composite moduli from the FWD were available. Regression analyses yielded correlation
261
coefficients ranging from 0.336 (Route 1A) to 0.950 (Witter Farm Road). In general,
correlation coefficients tended to increase as pavement thickness decreased. The data from
three test sites with asphalt thicknesses less than or equal to 127 mm (5 in.) were combined
and produced the best correlation with r2 = 0.873. Two test sites with an asphalt thickness of
152 mm (6 in.) followed with r2 = 0.559. However, when excluding unreasonably high
moduli greater than 4000 MPa, the correlation improves with r2 = 0.802. Route 1A was the
single test site with a 180 mm (7 in.) asphalt thickness and produced the poorest correlation
with r2 = 0.336. A regression analysis combining all asphalt surfaced test sites produced a
correlation coefficient of 0.531. Again, when moduli greater than 4000 MPa are excluded,
the correlation improved with r2 = 0.809. The average FWD and PFWD composite moduli
were lower for the 178 mm (7 in.) asphalt thickness than the 127 mm (5 in.) thickness. This
is contrary to expectations, since thicker pavements would be expected to yield higher
composite moduli.
Subbase moduli were derived from backcalculating traditional FWD data using
Evercalc. For each site, subbase moduli were plotted against composite moduli as measured
with the Prima 100 PFWD to determine if a correlation existed between the two measured
variables. Regression analyses yielded correlation coefficients ranging from 0.163 (Route
1A) to 0.807 (Knapp Airport Parking Lot). In general, correlation coefficients tended to
increase as pavement thickness decreased. Five test sites with an asphalt thicknesses equal to
127 mm (5 in.) had an r2 = 0.508. However, when excluding moduli greater than 5000 MPa,
the correlation improves with r2 = 0.693. One test site (Route 126) with a 150 mm (6 in.)
asphalt thickness produced the best correlation with r2 = 0.698. Route 1A served as the
single test site with a 180 mm (7 in.) asphalt thickness produced the poorest correlation with
262
r2 = 0.363. Regression analysis was completed for all the data combined, this yielded a
correlation coefficient of 0.485, however, when excluding all moduli greater than 5000 MPa
(Figure 4.59), the correlation improved with r2 = 0.654. These results suggest that the PFWD
composite modulus is influenced by the stiffness of the subbase.
6.1.3.4 Impact Stiffness Modulus
The Impact Stiffness Modulus (ISM) is the ratio of the applied force (kN) to the
deflection (μm) measured from the center sensor. Correlations between portable and
traditional FWD derived ISM tend to increase as pavement thickness decreases. This trend is
similar to that exhibited by composite and subbase moduli. Traditional and portable FWD
derived ISM are compared for three gravel surfaced test sites. Regression analyses yielded
correlation coefficients ranging from 0.638 (Lakeside Landing Road) to 0.914 (Crosstown
Road).
6.1.3.5 Comparison to Other Portable Devices
Loadman and Prima 100 PFWD composite moduli are compared to FWD derived
subbase moduli for two asphalt surfaced test sites in Rumney, New Hampshire. Correlations
developed indicate that for a given FWD derived subbase modulus, the Loadman PFWD
provides a composite modulus that is less than the corresponding value provided by the
Prima 100. The Prima 100 PFWD correlates better to FWD derived subbase moduli (r2 =
0.552) than composite moduli obtained from the Loadman PFWD (r2 = 0.245).
263
6.1.3.6 Evaluation of Field Testing Techniques
6.1.3.6.1 Loading Plate Diameter and Drop Weight
Operation of the Prima 100 PFWD with drop weights of 10, 15, and 20 kg (22, 33,
and 44 lb), and plate diameters of 100, 200, and 300 mm (4, 8, and 12 in.) were investigated.
In general, the 20 kg (44 lb) drop weight produced the lowest moduli and moduli that were
independent of loading plate diameter. The 15 kg (33 lb) weight produced moduli that were
greater than those obtained with the 20 kg (44 lb) and also did not vary significantly with
loading plate diameter. The highest moduli were obtained using the 10 kg (22 lb) drop
weight. At this weight, moduli decreased with increasing loading plate diameter. A
possible explanation for this behavior is that a small plate diameter and drop weight influence
only the upper portions of the pavement section and thus the deflection responses are
dominated by the stiffer pavement layer, producing a larger composite modulus. When plate
diameter and drop weight are increased, depth of influence is increased and the stiffness of
both the subbase and asphalt layers are reflected in the composite modulus, resulting in a
lower value.
6.1.3.6.2 Drop Height
Measurements were taken at three different drop heights at each test location
throughout the monitoring period. Drop heights used were approximately equal to 850, 630,
and 420 mm (33.5, 24.8, and 16.5 in.). In general, reduced drop heights produce moduli that
are slightly less than moduli derived from using the full (850 mm) drop height. This trend is
evident regardless of asphalt thickness; however, the differences tend to decrease with
increasing asphalt thickness. Decreasing drop height reduces the depth of influence. When
264
asphalt thickness increases and drop height is reduced simultaneously, measured moduli are
heavily influenced by the stiffness of the asphalt layer.
6.1.3.6.3 Moduli Derived From Additional Geophones
Three deflection sensors were used at each site to observe differences in moduli
derived from measurements taken from each of the geophones. Spacing of the sensors (as
measured from the center of the loading plate) is: 0, 207, and 407 mm (0, 8, and 16 in.). The
current Prima 100 software uses only the deflections of a single geophone to backcalculate
the modulus. The user selects which geophone will be used. Moduli derived from
measurements from the outer two geophones are significantly greater than the composite
moduli determined from the center geophone. Unless software can be developed to
incorporate the deflections from all three geophones simultaneously into a backcalculation
routine, the additional geophones provide little useful additional information.
6.1.3.6.4 Effect of Number of Drops on Composite Modulus
Six Prima 100 PFWD measurements were taken at each of three different drop
heights, at each test location. For the majority of points tested at the field sites; the first
measurement was less than subsequent measurements. This was consistent with observations
made by other researchers as discussed in the Literature Review. On the average, first drop
was less than the average of the remaining five drops by nearly 10%. However, the second
drop was only 1% less than the average of the remaining four drops. This shows that the
results of the first drop should always be neglected. It is recommended that the results from
drops two through six be averaged to obtain results that are representative of a test location.
265
6.1.3.6.5 Recommendations
Field testing techniques for monitoring seasonal stiffness variation in paved and
unpaved low volume roads using the Prima 100 PFWD were developed. The core of the
recommendations is that load restrictions are placed once the composite moduli measured
with the PFWD drops below 80% of the fully recovered baseline value measured during the
summer and early fall. The load restriction is then removed when the moduli recover to 80%
of the baseline value. The selection of 80% is arbitrary since the amount of damage that
would occur at the reduced modulus depends on individual pavement sections, allowable
vehicle weight, and traffic levels. Assessment of these factors was beyond the scope of this
study. Baseline value measurements and measurements during the spring thaw should be
made at the same locations. During the early portion of thawing period, it may be necessary
to take daily readings to monitor the sometimes rapid decrease in composite modulus. It is
recommended that the 300 mm (12 in.) loading plate, maximum drop height, and maximum
drop weight be used for testing.
6.1.4 Compaction Control
6.1.4.1 Laboratory Measurement Verification
As a check on the accuracy of the Nuclear Density Meter (NDM) used in the
laboratory tests, oven-dried water content and sand cone tests were also performed. Two
water content and one sand cone test were completed for each trial. Water contents
determined from both methods compared reasonably. A regression analysis yielded a
correlation coefficient of 0.55. As a result, using water contents determined from NDM
266
measurements was justified and were used for analysis and comparison. There was
essentially no correlation between the percent compaction NDM and sand cone. Moreover,
the sand cone predicted percent compactions that were greater than the NDM and many
results were in the range of 100% to 123%. Given the unreasonably high percent
compactions resulting from some of the sand cone results, it was concluded that the sand
cone results were unreliable and NDM derived values were used for comparison.
6.1.4.2 Field and Laboratory Test Results
Laboratory tests were performed on five soil types representative of New England
base and subbase aggregates. These materials include: one crushed material, one
construction sand, and three base/subbase aggregates. The field component included tests
on two subgrades, one construction sand product, two aggregates, and one reclaimed
stabilized base product.
For the laboratory tests, the composite moduli generally increased as percent
compaction increased. This was true for all samples with the exception of the New
Hampshire Gravel which exhibited the opposite trend. With the exception of the New
Hampshire Sand, the regression coefficients were less then 0.5 indicating poor correlation.
Combining all the results yielded a correlation coefficient of 0.045, indicating no correlation.
However, including only the results for Connecticut crushed gravel, New Hampshire sand,
and Wardwell gravel resulted in a higher correlation coefficient of 0.35, but still indicating a
poor correlation. Regression coefficients were slightly higher when separate correlations
were developed for samples dry and wet of optimum, however the r2 were still less than 0.5.
Results from the field test sites also indicate that as the degree of compaction increases,
267
composite modulus increases. In general, correlation coefficients were greater for field test
results compared to laboratory test results. Combining the results for the three base materials
tested in the field, resulted in a correlation coefficient of 0.818, which is a relatively strong
correlation. However, the significance of this correlation is diminished by the fact the water
content at all the field sites was dry of optimum. The Prima 100 PFWD also proved adequate
in measuring time dependent increases in composite modulus for asphalt stabilized base
products tested on Route 201 in The Forks, Maine, and at Commercial Recycling Systems in
Scarborough, Maine.
Laboratory results also show that there is a general trend that the composite moduli
tends to decrease as water content increases. Correlation coefficients ranged from 0.003
(Connecticut Crushed Gravel) to 0.814 (Wardwell Gravel). The low correlation coefficients
for several of the samples are due in part to the role that percent compaction plays in the
composite modulus, which is not accounted for when only water content is considered.
Combining at the laboratory results yielded a correlation coefficient of 0.285 which indicates
poor correlation. For measurements taken at field sites the correlation coefficient ranged
from 0.008 (Route 25 Gravel) to 0.521 (Route 25 Sand). However, water contents measured
at field sites were generally drier than -3% of the OWC and in some instances were as low as
-9% of the OWC, which are significantly different from water contents obtained during
laboratory tests.
Multivariable linear regression analyses (Neter, et al., 1982) were used to determine
the best fit line for composite modulus as a function of percent compaction and water
content. The R2 for the laboratory materials ranged from 0.141 (Connecticut crushed gravel)
to 0.867 (Wardwell gravel). Combining all laboratory samples produced an R2 of 0.326.
268
However, including only laboratory results for Connecticut crushed gravel, New Hampshire
sand, and Wardwell gravel increased the R2 to 0.624. This indicates that 62% of the
variation in composite modulus is explained by the percent compaction and water content
relative to optimum. The R2 for the field materials ranged from 0.001 (Route 25 gravel) to
0.679 (I-84 crushed gravel). Combining the three field sites where granular base was tested
yielded an R2 of 0.823, which indicates a reasonably strong correlation of composite modulus
with percent compaction and water content, independent of the type of material tested.
However, the water contents for the field sites were all dry of optimum which may limit the
significance of this result. The multi-variable linear regressions based on the three laboratory
samples indicated above and field results yielded predicted composite modulus at 95%
percent compaction that agreed within 20% which is reasonable agreement.
The results from two Prima 100 PFWDs were compared on gravel samples obtained
from Owen J. Folsom & Sons and Robert Wardwell & Sons Inc. The UMaine Prima 100
produced slightly higher moduli than the USFS Prima 100. The USFS PFWD was used first,
completing six measurements at three different drop heights, at each of the five test locations
before the process was repeated with the UMaine PFWD. The differences in moduli could,
in part, be due to additional compaction test points underwent as a result of using the USFS
PFWD first. A regression analysis yielded a correlation coefficient of 0.954. This shows
that two different Prima 100 units give results that correlate well with each other, even
though the modulus values are slightly different.
The Prima 100 PFWD and Clegg Impact Hammer (CIH) were used on all laboratory
samples. Prima 100 composite moduli were greater than composite moduli derived from
CIH measurements. The occurrence of a shallow bearing capacity failure caused by the
269
impact of the CIH could help to explain the large differences between moduli from the
devices. Additionally, the first drop with the CIH produced moduli that were less than those
derived from subsequent measurements. Finally, the CIH moduli tended to increase with
each subsequent drop.
The effect of operator technique was investigated by comparing the results obtained
by five different operators on one trial of Wardwell gravel (90% compaction, +10% wopt).
Average composite moduli for each user decreased as testing progressed which could
partially be caused by the presence of excess water in the material at the time of testing, as
minimal time was allotted for dissipation of pore water pressure between tests.
Recommendations were made for utilizing the Prima 100 PFWD as a tool to monitor
compaction. Tentative equivalences between percent compaction and composite modulus for
base course aggregates at optimum water content were provided. The recommendations
were based on a multi-variable linear regression on laboratory results for Connecticut
crushed gravel, New Hampshire sand, and Wardwell gravel. In addition, correction factors
to correct composite moduli measured at the field water content to the equivalent value at
optimum were proposed. Additional study on a wider range of base course aggregates and
water contents is recommended.
6.2 CONCLUSIONS
The conclusions listed below are based on the work presented in this report and the
experience of the researchers in using the Prima 100 PFWD.
270
6.2.1 Spring Thaw Monitoring
1. Prima 100 PFWD composite moduli follow similar trends to composite moduli and
subbase moduli as determined from FWD measurements on both asphalt and gravel
surfaced roads.
2. A strong correlation exists between portable and traditional FWD
composite moduli. The correlation increases with decreasing asphalt
thickness.
3. A reasonable correlation exists between Prima 100 PFWD composite
moduli and subbase moduli determined from FWD measurements on both
asphalt and gravel surfaced roads. The correlation increases with
decreasing pavement thickness.
4. Loadman PFWD provides a composite modulus which is greater than the
corresponding value provided by the Prima 100. The Prima 100 PFWD
correlates better than the Loadman PFWD to FWD subbase moduli.
5. The PFWD can be used as a tool to evaluate whether specific roadways
experience strength loss during the spring thaw and thus warrant load
restrictions. For roads where load restrictions are placed, the PFWD can be
used as an aid in determining when restrictions should be placed and removed.
6.2.2 Field Testing Techniques
1. Composite moduli increase with decreasing drop weight.
2. Composite moduli are independent of loading plate diameter when using
15 and 20 kg drop weight.
271
3. Reduced drop heights produce moduli that are slightly less than
moduli derived from using the full (850 mm) drop height.
4. Moduli derived from measurements from the outer two geophones are
significantly greater than composite moduli determined from the
center geophone.
5. The composite moduli determined from the first drop was generally less than
those derived from subsequent drops and should be ignored when computing
the composite modulus for a test location.
6.2.3 Compaction Control
1. Field and laboratory test results indicate that as percent compaction
increases, composite modulus increases. The correlation between the two
was poor, however, it was better for field tests compared to laboratory
tests.
2. Field and laboratory test results indicate that as water content increases,
composite modulus decreases. With the exception of one sample, the
correlation between the two variables was poor. The correlation was
better for field results compared to laboratory results.
3. A reasonable correlation exists between the PFWD composite modulus and the
combination of percent compaction and water content relative to optimum. These
results were used as the basis for a technique to use the PFWD as a tool for field
compaction control.
4. Two Prima 100 PFWDs produced nearly identical results.
272
5. Moduli determined from the Clegg Impact Hammer were significantly less
than those determined by the Prima 100. A marginal correlation exists
between the two devices.
6. Composite moduli did not appear to be affected by operator technique.
6.3 RECOMMENDATIONS FOR FURTHER RESEARCH
1. Additional spring thaw measurements should be made in order to verify the adequacy
of using 80% of the fully recovered composite modulus as a basis for placing and
removing spring load restrictions. Moreover, the influence of pavement structure,
allowable load, and traffic level on the adequacy of using 80% as the basis of placing
and removing load restrictions should be examined.
2. Software should be developed for the Prima 100 PFWD to incorporate deflections
from three geophones into a single backcalculation routine for determination of the
composite modulus.
3. Additional studies should be undertaken to better define the relationship between
composite modulus, percent compaction, and water content for a wide range of base
course aggregates, subgrade soils, and compaction water contents. These tests should
be based on field test results.
4. In situ measurements should be taken with the Prima 100 PFWD and compared with
laboratory determined resilient moduli.
273
REFERENCES
AASHTO T180, “Standard Method of Test for Moisture-Density Relations of Soils Using a 10 lb. (4.54 kg) Rammer and an 18 in. (457 mm) Drop,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
AASHTO T191, “Standard Method of Test for Density of Soil In-Place by the Sand-
Cone Method,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
AASHTO T238, “Standard Method of Test for Density of Soil and Soil-Aggregate in
Place by Nuclear Methods (Shallow Depth),” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
AASHTO T239, “Standard Method of Test for Moisture Content of Soil and Soil-
Aggregate in Place by Nuclear Methods (Shallow Depth),” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
AASHTO T265, “Standard Method of Test for Laboratory determination of Moisture
Content of Soils,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
AASHTO T311, “Standard Method of Test for Grain Size Analysis of Granular Soil
Materials,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, Part II, Tests, American Association of State Highway and Transportation Officials, Washington D.C.
Baumgardner, R.H. (1993). “Overview of permeable bases,” Materials: performance and
Prevention of Deficiencies and Failure, New York, FHWA-SA-94-045, pp. 275-287.
Bouchedid, M., and Humphrey, D.N. (2004). “Permeability of base material for Maine
roads,” report prepared for Maine Department of Transportation by Department of Civil and Environmental Engineering, University of Maine, Orono, Maine.
Brown, Warren (February 2003), Personal communication.
274
Canadian Strategic Highway Research Program, (2000). “Seasonal Load Restrictions in Canada and Around the World,” Technical Brief No. 21, September, 8pp.
Christensen, Ole R. (March 2003), Personal communication. Davies, Tom (1997). “Assessing the suitability of the ‘Loadman’ single point falling
weight deflectometer to tracking the change in strength in thin asphalt surfaced roads through spring thaw in Saskatchewan,” UNB International Symposium on Thin Pavements, Surface Treatments, and Unbound Roads, New Brunswick, Canada.
Dynatest International, (2004). http://www.dynatest.com/hardware/fwd_hwd.htm Federal Highway Administration (1994). “Pavement Deflection Analysis: Participant
Workbook,” Report No. FHWA-HI-94-021, NHI Course No. 13127, U.S. Department of Transportation, February 1994.
Fetten, C.P., and Humphrey, D.N. (1998). “Instrumentation and Performance of
Geosynthetics Beneath Flexible Pavements in Winterport and Frankfort, Maine,” A Study for the Maine Department of Transportation, by The Department of Civil and Environmental Engineering, University of Maine, Orono, Maine, pp. 137.
Fleming, P.R., and Rogers, C.D.F., (1995). “Assessment of Pavement Foundations during
Construction,” Transport, Proceedings of the I.C.E., vol. 111, No. 2, pp. 105-115. Fleming, P.R., Frost, M.W., and Rogers, C.D.F. (2000). “A comparison of devices for
measuring stiffness in-situ,” Unbound Aggregates in Road Construction, Proceedings of the Fifth International Symposium on Unbound Aggregates in Roads/UNBAR 5, Nottingham, United Kingdom, 21-23 June 2000, A.A. Balkema, Rotterdam, Netherlands, pp. 239-246.
Fleming, P.R., Lambert, J.P., and Frost, M.W. (2002). “In-situ assessment of stiffness
modulus for highway foundations during construction,” Proceedings of the Ninth International Conference on Asphalt Pavements, Copenhagen 2002, Conference Proceedings in PDF Format, Vol. 1, CD-ROM.
Groenendijk, J., Van Haasteren, C.R., and van Niekerk, A.A. (2000). “Comparison of
stiffness moduli of secondary road base materials under laboratory and in-situ conditions,” Unbound Aggregates in Road Construction, Proceedings of the Fifth International Symposium on Unbound Aggregates in Roads/UNBAR 5, Nottingham, United Kingdom, 21-23 June 2000, A.A. Balkema, Rotterdam, Netherlands, pp. 201-208.
Gros, Christophe (1993). “Use of a Portable Falling Weight Deflectometer; ‘Loadman’,”
Publications of Road and Transport Laboratory 20, University of Oulu, Finland.
275
Helstrom, C.L., and Humphrey, D.N. (2005). “ Performance and Effectiveness of a Thin Pavement Section Using Geogrids and Drainage Geocomposite In A Cold Region,” A Study for the New England Transportation Consortium, by The Department of Civil and Environmental Engineering, University of Maine, Orono, Maine, pp. 194.
Holtz, Robert D. and Kovacs, William D. (1981). An Introduction To Geotechnical
Engineering, N.M. Newmark and W.J. Hall, Prentice Hall, Englewood Cliffs, New Jersey.
Honkanen, Pentti (1997). “Loadman: Comparison of Test Results Obtained Using
Loadman with Those Obtained Using a Falling Weight Deflectometer and Plate Bearing Test,” Finnish National Road Administration, District Surveys and Studies, Turku 1991.
Huang, Yang H. (2004). Pavement Analysis and Design, Second Edition, Pearson
Education, Inc., Pearson Prentice Hall, Upper Saddle River, New Jersey. Janoo, Vincent, and Cortez, Edel, (2002). “Pavement Evaluation in Cold Regions,”
Proceedings of the Eleventh International Conference on Cold Regions Engineering. Anchorage, Alaska, 20-22 May 2002, ASCE.
Kamiura, M., Sekine, E., Abe, N., and Maruyama, T. (2000). “Stiffness evaluation of the
subgrade and granular aggregates using the portable FWD,” Unbound Aggregates in Road Construction, Proceedings of the Fifth International Symposium on Unbound Aggregates in Roads/UNBAR 5, Nottingham, United Kingdom, 21-23 June 2000, A.A. Balkema, Rotterdam, Netherlands, pp. 239-246.
Kestler, M.A., Knight, T., and Krat, A.S. (2000). “Thaw Weakening and Load
Restriction Practices on Low Volume Roads,” ERDC/CRREL TR-00-6, Technical Report.
Langfelder, L.J., and Nivargikar, V.R. (1967). “Some factors influencing shear strength
and compressibility of compacted soils,” Highway Research Board, No. 177, National Research Council, pp. 4-21.
Lawrence, B.K, Chen, L.H., and Humphrey, D.N. (2000). “Use of Tire Chip/Soil
Mixtures to Limit Frost Heave and Pavement Damage of Paved Roads,” A Study for the New England Transportation Consortium, by The Department of Civil and Environmental Engineering, University of Maine, Orono, Maine, pp. 316.
Lenke, L.R., McKeen, R.G., and Grush, M.P. (2003). “Laboratory Evaluation of
GeoGauge for Compaction Control,” Transportation Research Record 1849, Washington, D.C.
276
Livneh, Moshe (1997). “A portable FWD for determining in-situ asphalt layers moduli,” Mechanical Tests for Bituminous Materials, Proceedings of the Fifth International Rilem Symposium MTBM, Lyon, France, 14-16 May 1997.
Livneh, Moshe (1997). “Single-Measurement Estimation of In Situ Asphalt-Layer
Moduli with Portable Falling Weight Deflectometer,” Transportation Research Record 1570, Washington, D.C.
Livneh, M., Livneh, N.A., and Elhadad, E. (1997). “Determining a Pavement Modulus
from Portable FWD Testing,” Geotechnical Testing Journal, GTJODJ, Vol. 20, No. 4, December 1997, pp. 373-383.
Lukanen, E., Stubstad, R., and Briggs, R. (2000). “Temperature Predictions and
Adjustment Factors for Asphalt Pavement,” Report No. FHWA-RD-98-085, FHWA, U.S. Department of Transportation.
Nazzal, M. (2003). “Field Evaluation of in-situ test technology for Qc/QA during
construction of pavement layers and embankments,” M.S. Thesis, The Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, Louisiana, pp. 112.
Neter, J., Wasserman, W., and Whitmore, G.A. (1982). Applied Statistics, Second
Edition, Allyn and Bacon, Inc., Boston, Massachusetts. Newcomb, D.E., and Birgisson, B. (1999). “Measuring In Situ Mechanical properties of
Pavement Subgrade Soils,” NCHRP synthesis 278, Transportation Research Board, National Research Council, Washington, D.C., 1999.
Ovik, J.M., Siekmeier, J.A., and Van Deusen, D.A. (2000). “Improved Spring Load
Restriction Guidelines Using Mechanistic Analysis,” Minnesota Department of Transportation, Report No. MN/RC-2000-18, 87 pp.
Pidwerbesky, B. (1997). “Predicting rutting in unbound granular basecourses from
Loadman and other in situ non-destructive tests,” Road and Transport Research, Volume 6 No. 3, September.
Pidwerbesky, B. (1997). “Evaluation of non-destructive in situ tests for unbound granular
pavements,” IPENZ Transactions, Vol. 24, No. 1/CE, 1997. Rutherford, M.S., Mahoney, J.P., Hicks, R.G., and Rwebingira, T. (1985). “Guidelines
for Spring Highway Use Restrictions,” Report No. FHWA-RD-86-501, Washington State Department of Transportation, Seattle, WA, 1985.
277
Siekmeier, J.A., Young, D., and Beberg, D., (2000). ”Comparison of the Dynamic Cone Penetrometer with Other Tests During Subgrade and Granular Base Characterization in Minnesota,” Nondestructive Testing of Pavements and Backcalculation of Moduli, Third Volume, ASTM STP 1375, S.D. Tayabji and E.O. Lukanen, Eds., American Society for Testing and Materials, West Conshohocken, PA, 2000.
Thom, N.H., and Fleming, P.R. (2002). “Experimental and Theoretical Comparison of
Dynamic Plate Testing Methods,” Bearing Capacity of Roads, Railways and Airfields, Proceedings of the Sixth International Conference on the Bearing Capacity of Roads and Airfields, Lisbon, Portugal, 24-26 June 2002, Correia, A.G., Vol.1, pp. 731-740.
Thompson, Marshall and Garcia, Gabriel (2003). “Subgrade strength/stiffness
evaluation” Draft White Paper 1, Department of Civil Engineering, University of Illinois Urbana-Champaign.
Van Deusen, Dave (1998). “Improved Spring Load Restriction Guidelines Using
Mechanistic Analysis,” Proceedings of the Ninth International Conference on Cold Regions Engineering. Duluth, Minnesota, 27-30 September 1998, ASCE.
Van Gurp, C., Groenendijk, J., and Beuving, E. (2000). “Experience with various types of
foundation tests,” Unbound Aggregates in Road Construction, Proceedings of the Fifth International Symposium on Unbound Aggregates in Roads/UNBAR 5, Nottingham, United Kingdom, 21-23 June 2000, A.A. Balkema, Rotterdam, Netherlands, pp. 239-246.
Walpole, R.E., and Myers, R.H. (1978), Probability and Statistics for Engineers and
Scientists, MacMillan Publishing, New York, 580 pp. Washington State Department of Transportation (1999). “WSDOT Pavement Guide,”
Volume 3, Seattle, Washington. Whaley, Andrew M. (1994). “Non-Destructive Pavement Testing Equipment: Loadman,
Falling Weight Deflectometer, Benkelman Beam, Clegg Hammer,” Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand, 19 pp.
278
(BLANK PAGE)
279
APPENDIX A -
GRAIN SIZE DISTRIBUTION CURVES
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.1 Grain size distribution of Connecticut crushed gravel.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.2 Grain size distribution of New Hampshire sand.
280
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100PE
RC
EN
T P
ASS
ING
Figure A.3 Grain size distribution of New Hampshire gravel.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.4 Grain size distribution of bottom 1 ft of OJF gravel.
281
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.5 Grain size distribution of Owen J. Folsom gravel.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.6 Grain size distribution of Wardwell gravel.
282
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100PE
RC
EN
T P
ASS
ING
Figure A.7 Grain size distribution of crushed gravel tested at I-84, Southington, Connecticut.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.8 Grain size distribution of existing subgrade material tested at I-84, Southington, Connecticut.
283
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100PE
RC
EN
T P
ASS
ING
Figure A.9 Grain size distribution of construction sand tested at Route 25, Effingham/Freedom, New Hampshire.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.10 Grain size distribution of gravel tested at Route 25, Effingham/Freedom, New Hampshire.
284
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.11 Grain size distribution of MaineDOT Type D gravel tested at Route 26, New Gloucester, Maine.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.12 Grain size distribution of MaineDOT Type E gravel tested at Route 26, New Gloucester, Maine.
285
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01
GRAIN DIAMETER (mm)
10
20
30
40
50
60
70
80
90
100PE
RC
EN
T P
ASS
ING
Figure A.13 Grain size distribution of existing subgrade tested at CPR, Scarborough,
Maine.
1E+003 1E+002 1E+001 1 0.1 0.01 0.001
GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Test Section 3 - STA 1+670Test Section 3 - STA 1+670
BOULDERS GRAVEL SAND FINES
Figure A.14 Grain size distribution of existing subgrade material at Route 126 (Section 3), Monmouth/Litchfield, Maine.
286
1E+003 1E+002 1E+001 1 0.1 0.01 0.001GRAIN DIAMETER (mm)
0
10
20
30
40
50
60
70
80
90
100PE
RC
EN
T P
ASS
ING
Test Section 8 - STA 4+040 - Depth = 1.52 to 1.83mTest Section 8 - STA 4+040 - Depth = 0.52 to 1.52mTest Section 8 - STA 4+040 - Depth = 1.83 to 3.00m
BOULDERS GRAVEL SAND FINES
Figure A.15 Grain size distribution of existing subgrade material at Route 126 (Section 8), Monmouth/Litchfield, Maine.
GRAVEL SAND FINESBOULDERS
1E+003 100 10 1 0.1 0.01GRAIN DIAMETER (mm)
10
20
30
40
50
60
70
80
90
100
PER
CE
NT
PA
SSIN
G
Figure A.16 Grain size distribution of existing subbase material at Stinson Lake Road, Rumney, New Hampshire.
287
Figure A.17 Grain size distribution of tire chip / soil mixtures at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000).
Figure A.18 Grain size distribution of MaineDOT Type D subbase used at Witter Farm Road, Orono, Maine (Lawrence, et al., 2000).
288
Figure A.19 Grain size distribution of MaineDOT Type D subbase used at Witter Farm
Road, Orono, Maine (Lawrence, et al., 2000).
Figure A.20 Grain size distribution of subbase material at Route 1A, Frankfort/Winterport,
Maine (Fetten and Humphrey, 1998).
289
APPENDIX B -
MOISTURE DENSITY CURVES
2 4 6 8 10Water Content (%)
2.2
2.24
2.28
2.32
Dry
Den
sity
(Mg/
m3 )
138
140
142
144
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.31 Mg/m3 (144 lb/ft3)
wop
t = 7
.4 %
Figure B.1 Connecticut crushed gravel moisture density curve.
6 8 10 12 14 16 18Water Content (%)
1.8
1.85
1.9
1.95
2
2.05
2.1
Dry
Den
sity
(Mg/
m3 )
116
120
124
128
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.06 Mg/m3 (128 lb/ft3)
wop
t = 1
0.6
%
Figure B.2 New Hampshire sand moisture density curve
290
4 8 12 16 20Water Content (%)
1.8
1.9
2
2.1
Dry
Den
sity
(Mg/
m3 )
116
120
124
128
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.05 Mg/m3 (128 lb/ft3)
wop
t = 9
.2 %
Figure B.3 New Hampshire gravel moisture density curve.
4 6 8 10 12Water Content (%)
1.88
1.92
1.96
2
Dry
Den
sity
(Mg/
m3 )
120
122
124
126
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 1.998 Mg/m3 (125 lb/ft3)
wop
t = 1
0.5
%
Figure B.4 Owen J. Folsom #1 moisture density curve.
291
4 6 8 10 12 14Water Content (%)
1.84
1.88
1.92
1.96
2
2.04
Dry
Den
sity
(Mg/
m3 )
116
120
124
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.00 Mg/m3 (125 lb/ft3)
wop
t = 1
1.2
%
Figure B.5 Owen J. Folsom #2 moisture density curve.
2 4 6 8 10 12Water Content (%)
2
2.04
2.08
2.12
Dry
Den
sity
(Mg/
m3 )
126
128
130
132
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.099 Mg/m3 (131 lb/ft3)
wop
t = 5
.1 %
Figure B.6 Wardwell gravel moisture density curve.
292
6 8 10 12 14Water Content (%)
2.08
2.1
2.12
2.14
2.16
2.18
Dry
Den
sity
(Mg/
m3 )
130
132
134
136
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.17 Mg/m3 (135 lb/ft3)
wop
t = 1
1.4
%
Figure B.7 Moisture density curve of sand tested at Route 25, Effingham/Freedom, New Hampshire.
6 8 10 12 14Water Content (%)
1.86
1.88
1.9
1.92
1.94
Dry
Den
sity
(Mg/
m3 )
117
118
119
120
121
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 1.92 Mg/m3 (120 lb/ft3)
wop
t = 1
2.3
%
Figure B.8 Moisture density curve of gravel tested at Route 25, Effingham/Freedom, New Hampshire.
293
6 8 10 12 14 16Water Content (%)
1.88
1.92
1.96
2
Dry
Den
sity
(Mg/
m3 )
118
120
122
124
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 1.99 Mg/m3 (124 lb/ft3)
wop
t = 1
2.0
%
Figure B.9 Moisture density curve of MaineDOT Type D at Route 26, New Gloucester, Maine.
6 8 10 12 14Water Content (%)
1.86
1.88
1.9
1.92
1.94
Dry
Den
sity
(Mg/
m3 )
117
118
119
120
121
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 1.94 Mg/m3 (121 lb/ft3)
wop
t = 1
0.8
%
Figure B.10 Moisture density curve of MaineDOT Type E at Route 26, New Gloucester, Maine.
294
2 4 6 8 10 12Water Content (%)
1.98
2
2.02
2.04
2.06
Dry
Den
sity
(Mg/
m3 )
124
125
126
127
128
Dry
Den
sity
(lb/
ft3 )
γdry(max) = 2.05 Mg/m3 (128 lb/ft3)
wop
t = 5
.5 %
Figure B.11 Moisture density curve of existing subgrade material at CPR, Scarborough, Maine.
295
APPENDIX C -
COMPACTION CONTROL RAW LABORATORY DATA
Table C.1 Summary of Connecticut crushed gravel raw laboratory data.
Target Test
Test Point
Dry Density Mg/m3 (lb/ft3)
Percent Compaction
(%)
Water Content
(%)
Water Content
Relative to Optimum
(%)
PFWD CompositeModulus
(MPa)
1 2.07 (129) 90 5.6 -1.8 90 2 2.00 (125) 87 6.3 -1.1 86 3 2.00 (125) 87 6.1 -1.3 115 4 2.13 (133) 93 5.8 -1.6 97 90
%, w
opt
5 2.10 (131) 91 5.2 -2.2 99 1 2.24 (140) 97 5.6 -1.8 130 2 2.26 (141) 98 6.6 -0.8 114 3 2.24 (140) 97 6.5 -0.9 202 4 2.23 (139) 97 5.7 -1.7 170 95
%, w
opt
5 2.27 (142) 99 5.4 -2.0 140 1 2.37 (148) 103 6.6 -0.8 152 2 2.35 (147) 102 6.3 -1.1 140 3 2.34 (146) 101 6.7 -0.7 184 4 2.39 (149) 104 6.7 -0.7 219 10
0%, w
opt
5 2.40 (150) 104 6.3 -1.1 172 1 2.35 (147) 102 6.1 -1.3 122 2 2.26 (141) 98 6.1 -1.3 114 3 2.31 (144) 100 5.4 -2.0 121 4 2.34 (146) 101 5.1 -2.3 137 95
%,
+3%
wop
t
5 2.32 (145) 101 4.7 -2.7 130 1 2.21 (138) 96 4.4 -3.0 132 2 2.11 (132) 92 4.1 -3.3 146 3 2.10 (131) 91 4.4 -3.0 199 4 2.15 (134) 93 4.5 -2.9 176 95
%,
-3%
wop
t
5 2.15 (134) 93 4.2 -3.2 194
296
Table C.2 Summary of New Hampshire sand raw laboratory data.
Target Test
Test Point
Dry Density Mg/m3 (lb/ft3)
Percent Compaction
(%)
Water Content
(%)
Water Content
Relative to Optimum
(%)
PFWD CompositeModulus
(MPa)
1 1.76 (110) 86 9.0 -1.6 49 2 1.78 (111) 86 8.4 -2.2 50 3 1.79 (112) 87 7.8 -2.8 60 4 1.79 (112) 87 7.2 -3.4 58 90
%, w
opt
5 1.76 (110) 86 8.5 -2.1 72 1 1.89 (118) 92 8.2 -2.4 77 2 1.87 (117) 91 8.5 -2.1 81 3 1.83 (114) 89 8.3 -2.3 85 4 1.86 (116) 90 7.8 -2.8 94 5 1.89 (118) 92 8.2 -2.4 88 1 1.91 (119) 93 9.3 -1.3 122 2 1.87 (117) 91 9.5 -1.1 115 3 1.86 (116) 91 8.5 -2.1 112 4 1.91 (119) 93 8.7 -1.9 125 5 1.94 (121) 94 8.4 -2.2 117 1 1.91 (119) 93 9.4 -1.2 158 2 1.87 (117) 93 9.7 -0.9 122 3 1.86 (116) 93 8.7 -1.9 151 4 1.91 (119) 94 8.9 -1.7 170
95%
, wop
t
5 1.94 (121) 95 8.5 -2.1 120 1 1.91 (119) 93 9.3 -1.3 95 2 1.94 (121) 95 8.6 -2.0 107 3 1.94 (121) 95 8.4 -2.2 116 4 1.91 (119) 93 9.3 -1.3 130 5 1.95 (122) 95 8.2 -2.4 114 1 1.87 (117) 91 12.7 +2.1 60 2 1.86 (116) 91 12.7 +2.1 50 3 1.87 (117) 91 10.9 +0.3 68 4 1.89 (118) 92 10.4 -0.2 64
95%
, +3%
wop
t
5 1.84 (115) 90 11.7 +1.1 78 1 1.91 (119) 93 8.6 -2.0 172 2 1.94 (121) 95 7.8 -2.8 160 3 1.92 (120) 93 7.8 -2.8 139 4 1.92 (120) 94 8.3 -2.3 179 95%
, -3%
wop
t
5 1.94 (121) 95 7.7 -2.9 165
297
Table C.3 Summary of New Hampshire gravel raw laboratory data.
Target Test
Test Point
Dry Density Mg/m3 (lb/ft3)
Percent Compaction
(%)
Water Content
(%)
Water Content
Relative to Optimum
(%)
PFWD CompositeModulus
(MPa)
1 1.94 (121) 94 12.7 +3.5 68 2 1.92 (120) 94 11.6 +2.4 73 3 1.97 (123) 96 11.3 +2.1 72 4 2.02 (126) 99 11.6 +2.4 69 90
%, w
opt
5 1.97 (124) 96 10.7 +1.5 79 1 1.92 (120) 93 11.8 +2.6 100 2 1.94 (121) 95 10.9 +1.7 89 3 1.94 (121) 94 11.4 +2.2 96 4 1.92 (120) 94 10.6 +1.4 93 95
%, w
opt
5 2.02 (126) 99 9.7 +0.5 104 1 2.03 (127) 99 8.6 -0.6 90 2 2.00 (125) 98 7.4 -1.8 82 3 2.00 (125) 98 7.1 -2.1 136 4 2.02 (126) 99 8.1 -1.1 103 10
0%, w
opt
5 1.97 (124) 97 7.8 -1.4 121 1 2.05 (128) 100 9.6 +0.4 70 2 2.03 (127) 99 10.7 +1.5 69 3 2.00 (125) 98 9.3 +0.1 98 4 1.97 (123) 96 9.6 +0.4 87 95
%,
+3%
wop
t
5 2.07 (129) 101 8.4 -0.8 80 1 2.02 (126) 98 7.7 -1.5 96 2 1.79 (112) 88 7.3 -1.9 99 3 1.92 (120) 94 7.1 -2.1 131 4 1.89 (118) 92 8.9 -0.3 130 95
%,
-3%
wop
t
5 1.84 (115) 90 7.5 -1.7 150
298
Table C.4 Summary of Owen J. Folsom gravel raw laboratory data.
Target Test
Test Point
Dry Density Mg/m3 (lb/ft3)
Percent Compaction
(%)
Water Content
(%)
Water Content
Relative to Optimum
(%)
PFWD CompositeModulus
(MPa)
1 1.94 (121) 97 12.7 +1.5 49 2 1.92 (120) 96 11.6 +0.4 51 3 1.97 (123) 99 11.3 +0.1 55 4 2.02 (126) 101 11.6 +0.4 58 90
%, w
opt
5 1.99 (124) 99 10.7 -0.5 60 1 1.91 (119) 95 11.8 +0.6 59 2 1.94 (121) 97 10.9 -0.3 60 3 1.94 (121) 97 11.4 +0.2 63 4 1.92 (120) 96 10.6 -0.6 65 95
%, w
opt
5 2.02 (126) 101 9.7 -1.5 64 1 2.02 (126) 100 11.2 0.0 69 2 1.97 (123) 99 11.6 +0.4 85 3 2.05 (128) 103 10.4 -0.8 103 4 2.02 (126) 101 10.7 -0.5 89 10
0%, w
opt
5 2.00 (125) 100 9.0 -2.2 88 1 1.95 (122) 98 13.5 +2.3 56 2 2.03 (127) 102 12.4 +1.2 61 3 2.08 (130) 104 9.2 -2.0 71 4 1.99 (124) 99 10.9 -0.3 61 95
%,
+3%
wop
t
5 2.00 (125) 100 11.1 -0.1 68 1 2.02 (126) 100 7.7 -3.5 77 2 1.79 (112) 90 7.3 -3.9 72 3 1.92 (120) 96 7.1 -4.1 86 4 1.89 (118) 94 8.9 -2.3 79 95
%,
-3%
wop
t
5 1.84 (115) 92 7.5 -3.7 76
299
Table C.5 Summary of Wardwell gravel raw laboratory data.
Target Test
Test Point
Dry Density Mg/m3 (lb/ft3)
Percent Compaction
(%)
Water Content
(%)
Water Content
Relative to Optimum
(%)
PFWD CompositeModulus
(MPa)
1 1.97 (124) 95 6.8 +1.7 99 2 2.02 (126) 96 5.6 +0.5 87 3 1.95 (122) 93 6.2 +1.1 85 4 1.81 (113) 86 5.9 +0.8 107 90
%, w
opt
5 1.92 (120) 92 5.4 +0.3 90 1 2.05 (128) 97 5.6 +0.5 119 2 2.08 (130) 99 6.1 +1.0 118 3 2.00 (125) 95 6.3 +1.2 135 4 2.03 (127) 97 6.1 +1.0 143 95
%, w
opt
5 2.13 (133) 101 6.8 +1.7 126 1 2.07 (129) 99 7.3 +2.2 126 2 2.16 (135) 103 6.2 +1.1 121 3 2.15 (134) 102 6.5 +1.4 141 4 2.11 (132) 100 6.7 +1.6 132 10
0%, w
opt
5 2.15 (134) 102 6.6 +1.5 115 1 1.89 (118) 90 8.4 +3.3 68 2 2.03 (127) 97 7.4 +2.3 72 3 1.81 (113) 86 8.6 +3.5 72 4 1.84 (115) 88 7.8 +2.7 76 5 1.94 (124) 94 8.3 +3.2 79 1 1.97 (123) 94 11.5 +6.4 56 2 1.92 (120) 92 11.1 +6.0 56 3 1.94 (121) 92 11.1 +6.0 58 4 1.97 (123) 94 11.3 +6.2 62 5 1.94 (121) 93 10.5 +5.4 63 1 1.87 (117) 89 15.1 +10.0 12 2 1.91 (119) 91 14.4 +9.3 19 3 1.87 (117) 89 14.1 +9.0 15 4 1.94 (124) 94 13.9 +8.8 16
95%
, +3%
wop
t
5 1.86 (116) 88 18.9 +13.8 13 1 1.97 (123) 94 4.3 -0.8 127 2 2.11 (132) 101 3.7 -1.4 144 3 2.10 (131) 100 4.1 -1.0 179 4 1.91 (119) 91 4.1 -1.0 148 95
%, -
3%w
opt
5 2.07 (129) 99 4.0 -1.1 144